Jun Suzuki

J-GLOBAL         Last updated: Nov 17, 2018 at 09:06
 
Avatar
Name
Jun Suzuki
URL
http://www.cl.ecei.tohoku.ac.jp/~jun/
Affiliation
Tohoku University
Section
Graduate School of Information Sciences
Job title
Associate Professor

Research Areas

 
 

Academic & Professional Experience

 
Apr 2018
 - 
Today
Tohoku University
 
Apr 2017
 - 
Today
Visiting Resercher, Center for Advanced Intelligence Project, RIKEN
 
Apr 2018
 - 
Today
Research Professor, Communication Science Laboratories, Nippon Telegraph and Telephone Corporation
 
Apr 2001
 - 
Mar 2018
Communication Science Laboratories, Nippon Telegraph and Telephone Corporation
 

Published Papers

 
Shun Kiyono,Jun Suzuki,Kentaro Inui
AAAI-2019 (To appear)   abs/1810.05788    2019   [Refereed]
Sho Yokoi,Sosuke Kobayashi,Kenji Fukumizu,Jun Suzuki,Kentaro Inui
EMNLP-2018   abs/1809.00800 1763-1775   Nov 2018   [Refereed]
Jun Suzuki,Sho Takase,Hidetaka Kamigaito,Makoto Morishita,Masaaki Nagata
Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, ACL 2018, Melbourne, Australia, July 15-20, 2018, Volume 2: Short Papers   612-618   2018   [Refereed]
Makoto Morishita,Jun Suzuki,Masaaki Nagata
Proceedings of the 27th International Conference on Computational Linguistics, COLING 2018, Santa Fe, New Mexico, USA, August 20-26, 2018   618-629   2018   [Refereed]
Motoki Sato,Jun Suzuki,Hiroyuki Shindo,Yuji Matsumoto
Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, IJCAI 2018, July 13-19, 2018, Stockholm, Sweden.   abs/1805.02917 4323-4330   2018   [Refereed]
Sho Takase,Jun Suzuki,Masaaki Nagata
Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussels, Belgium, October 31 - November 4, 2018   abs/1808.10143 4599-4609   2018   [Refereed]
Makoto Morishita,Jun Suzuki,Masaaki Nagata
Proceedings of the Third Conference on Machine Translation: Shared Task Papers, WMT 2018, Belgium, Brussels, October 31 - November 1, 2018   461-466   2018   [Refereed]
Naoki Mizukami,Jun Suzuki,Hirotaka Kameko,Array
Advances in Computer Games - 15th International Conferences, ACG 2017, Leiden, The Netherlands, July 3-5, 2017, Revised Selected Papers   165-175   2017   [Refereed]
Makoto Morishita,Jun Suzuki,Masaaki Nagata
Proceedings of the 4th Workshop on Asian Translation, WAT@IJCNLP 2017, Taipei, Taiwan, November 27- December 1, 2017   89-94   2017   [Refereed]
Jun Suzuki,Masaaki Nagata
Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017, Valencia, Spain, April 3-7, 2017, Volume 2: Short Papers   abs/1701.00138 291-297   2017   [Refereed]
Hirotaka Kameko,Jun Suzuki,Naoki Mizukami,Array
Computer Games - 6th Workshop, CGW 2017, Held in Conjunction with the 26th International Conference on Artificial Intelligence, IJCAI 2017, Melbourne, VIC, Australia, August, 20, 2017, Revised Selected Papers   46-60   2017   [Refereed]
Itsumi Saito,Jun Suzuki,Kyosuke Nishida,Kugatsu Sadamitsu,Satoshi Kobashikawa,Ryo Masumura,Yuji Matsumoto 0001,Junji Tomita
Proceedings of the Eighth International Joint Conference on Natural Language Processing, IJCNLP 2017, Taipei, Taiwan, November 27 - December 1, 2017, Volume 2: Short Papers   257-262   2017   [Refereed]
Shun Kiyono,Sho Takase,Jun Suzuki,Naoaki Okazaki,Kentaro Inui,Masaaki Nagata
CoRR   abs/1712.08302    2017   [Refereed]
Sho Takase,Jun Suzuki,Masaaki Nagata
Proceedings of the Eighth International Joint Conference on Natural Language Processing, IJCNLP 2017, Taipei, Taiwan, November 27 - December 1, 2017, Volume 2: Short Papers   abs/1709.08907 43-48   2017   [Refereed]
Tsutomu Hirao,Jun Suzuki,Masaaki Nagata,Masaaki Nishino
Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2017, Valencia, Spain, April 3-7, 2017, Volume 1: Long Papers   386-396   2017   [Refereed]
Masaaki Nishino,Jun Suzuki,Masaaki Nagata
Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, ACL 2016, August 7-12, 2016, Berlin, Germany, Volume 2: Short Papers      2016   [Refereed]
Sho Takase,Jun Suzuki,Naoaki Okazaki,Tsutomu Hirao,Masaaki Nagata
Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, EMNLP 2016, Austin, Texas, USA, November 1-4, 2016   1054-1059   2016   [Refereed]
Jun Suzuki,Masaaki Nagata
Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, IJCAI 2016, New York, NY, USA, 9-15 July 2016   2046-2052   2016   [Refereed]
Jun Suzuki,Masaaki Nagata
NAACL HLT 2016, The 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, San Diego California, USA, June 12-17, 2016   1145-1151   2016   [Refereed]
Nishino Masaaki, Suzuki Jun, Umetani Shunji, Hirao Tsutomu, Nagata Masaaki
Journal of Natural Language Processing   23(2) 175-194   2016
Sequence alignment, which involves aligning elements of two given sequences, occurs in many natural language processing (NLP) tasks such as sentence alignment. Previous approaches for solving sequence alignment problems in NLP can be categorized i...
Tsutomu Hirao,Masaaki Nishino,Yasuhisa Yoshida,Jun Suzuki,Norihito Yasuda,Masaaki Nagata
IEEE/ACM Trans. Audio, Speech & Language Processing   23(11) 2081-2092   2015   [Refereed]
Jun Suzuki,Masaaki Nagata
Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing of the Asian Federation of Natural Language Processing, ACL 2015, July 26-31, 2015, Beijing,   186-191   2015   [Refereed]
Jun Suzuki,Masaaki Nagata
Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligence, July 27 -31, 2014, Québec City, Québec, Canada.   1593-1599   2014   [Refereed]
Yasuhisa Yoshida,Jun Suzuki,Tsutomu Hirao,Masaaki Nagata
Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014, October 25-29, 2014, Doha, Qatar, A meeting of SIGDAT, a Special Interest Group of the ACL   1834-1839   2014   [Refereed]
Yotaro Kubo,Jun Suzuki,Takaaki Hori,Atsushi Nakamura
INTERSPEECH 2014, 15th Annual Conference of the International Speech Communication Association, Singapore, September 14-18, 2014   1068-1072   2014   [Refereed]
Akinori Fujino,Jun Suzuki,Tsutomu Hirao,Hisashi Kurasawa,Katsuyoshi Hayashi
Proceedings of the 11th NTCIR Conference on Evaluation of Information Access Technologies, NTCIR-11, National Center of Sciences, Tokyo, Japan, December 9-12, 2014      2014   [Refereed]
Hayashi Katsuhiko, Sudoh Katsuhito, Tsukada Hajime, Suzuki Jun, Nagata Masaaki
Journal of Natural Language Processing   21(5) 1037-1057   2014
This paper introduces a novel word re-ordering model for statistical machine translation that employs a shift-reduce parser for inversion transduction grammars. The proposed model also solves article generation problems simultaneously with word re...
数原 良彦, 鈴木 潤, 鷲崎 誠司
情報処理学会論文誌データベース(TOD)   6(2) 51-60   Mar 2013
ウェブスパム判別においては,あらかじめラベル付けされた訓練データを用いて機械学習の枠組みでスパム判別器を生成する方法が広く用いられている.本稿では,ウェブスパム判別において特に課題となる偽陽性率に着目し,偏りのある訓練データを用いた場合においても偽陽性率を抑えつつ,高精度な判別が可能となるマージン識別器のオンライン学習手法を提案する.提案手法では学習時にスパムと非スパム側に異なるマージンサイズを設定することで偽陽性率を抑え,クラスを確率的に選択したうえで当該クラスにおいて最大損失を与える事...
Jun Suzuki,Masaaki Nagata
Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, ACL 2013, 4-9 August 2013, Sofia, Bulgaria, Volume 2: Short Papers   18-23   2013   [Refereed]
Masaaki Nishino,Norihito Yasuda,Tsutomu Hirao,Jun Suzuki,Masaaki Nagata
Advances in Information Retrieval - 35th European Conference on IR Research, ECIR 2013, Moscow, Russia, March 24-27, 2013. Proceedings   772-775   2013   [Refereed]
Katsuhiko Hayashi,Katsuhito Sudoh,Hajime Tsukada,Jun Suzuki,Masaaki Nagata
Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, EMNLP 2013, 18-21 October 2013, Grand Hyatt Seattle, Seattle, Washington, USA, A meeting of SIGDAT, a Special Interest Group of the ACL   1382-1386   2013   [Refereed]
Katsuhito Sudoh,Jun Suzuki,Hajime Tsukada,Masaaki Nagata,Sho Hoshino,Yusuke Miyao
Proceedings of the 10th NTCIR Conference on Evaluation of Information Access Technologies, NTCIR-10, National Center of Sciences, Tokyo, Japan, June 18-21, 2013      2013   [Refereed]
Norihito Yasuda,Masaaki Nishino,Tsutomu Hirao,Jun Suzuki,Ryoji Kataoka
23rd International Workshop on Database and Expert Systems Applications, DEXA 2012, Vienna, Austria, September 3-7, 2012   126-130   2012   [Refereed]
鈴木 潤, 磯崎 秀樹, 永田 昌明
情報処理学会論文誌   52(11) 3038-3051   Nov 2011
係り受け解析では,正解係り受け構造が付与されたデータを用いた教師あり学習により解析器を学習するのが現在最も一般的な方法であり,データ量が十分あれば非常に高い解析精度が得られることが実証されている.しかし,さらなる解析精度向上のため,正解データを増やし続けるのは作成に要する費用や時間の観点で現実的な方策ではない.そこで本論文では,正解係り受け構造が付与されていないデータも利用して解析精度を向上させる,いわゆる半教師あり学習に基づく係り受け解析モデルとその学習法を提案する.実験では,係り受け解...
SUZUKI Jun, ISOZAKI Hideki
The IEICE transactions on information and systems   94(5) 908-918   May 2011
サンプル間に依存関係があるデータに対して大域的な最適化による識別学習を行うモデルとして,条件付確率場が提案され多くの実タスクで良好な性能を示している.条件付確率場のパラメータ推定(学習)は,確率場全体のゆう度,あるいは,事後確率に基づく目的関数を最大化する方法が一般的である.しかし,実タスクを評価する際に用いる評価指標は,ゆう度や事後確率でなく,タスクの目的に合わせてF値等の様々な評価関数が用いられる.そのために,タスクの評価指標と学習時の目的関数間にはしばしば不整合が起きることがある.し...
Jun Suzuki,Hideki Isozaki,Masaaki Nagata
The 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, Proceedings of the Conference, 19-24 June, 2011, Portland, Oregon, USA - Short Papers   636-641   2011   [Refereed]
Jun Suzuki,Kevin Duh,Masaaki Nagata
Fifth International Joint Conference on Natural Language Processing, IJCNLP 2011, Chiang Mai, Thailand, November 8-13, 2011   649-657   2011   [Refereed]
平尾 努, 鈴木 潤, 磯崎 秀樹
情報処理学会論文誌データベース(TOD)   2(1) 1-9   Mar 2009
従来の文短縮手法の多くは,入力された文を構文木として表現し,その部分木を削除することで,短縮文を生成する.このようなアプローチは文法的な短縮文を生成するという観点からは理にかなっている.しかし,多くの場合,人間は構文木の刈り込みだけで短縮文を生成するわけではない.これは,構文情報に過度に依存することが,高品質な文短縮を行うための妨げとなることを示している.そこで,本稿では,構文情報を用いない文短縮手法を提案する.短縮文の言語としてのもっともらしさを構文情報を用いずに評価するため,原文と大規...
Tsutomu Hirao,Jun Suzuki,Hideki Isozaki
ACL 2009, Proceedings of the 47th Annual Meeting of the Association for Computational Linguistics and the 4th International Joint Conference on Natural Language Processing of the AFNLP, 2-7 August 2009, Singapore   826-833   2009   [Refereed]
Jun Suzuki,Hideki Isozaki,Xavier Carreras,Michael Collins 0001
Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, EMNLP 2009, 6-7 August 2009, Singapore, A meeting of SIGDAT, a Special Interest Group of the ACL   551-560   2009   [Refereed]
Hirao Tsutomu, Suzuki Jun, Isozaki Hideki
Transactions of the Japanese Society for Artificial Intelligence   24(2) 223-231   2009
We derived the oracle summary with the highest ROUGE score that can be achieved by integrating sentence extraction with sentence compression from the reference abstract. The analysis results of the oracle revealed that summari...
Jun Suzuki,Hideki Isozaki
ACL 2008, Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics, June 15-20, 2008, Columbus, Ohio, USA   665-673   2008   [Refereed]
Akinori Fujino,Hideki Isozaki,Jun Suzuki
Third International Joint Conference on Natural Language Processing, IJCNLP 2008, Hyderabad, India, January 7-12, 2008   823-828   2008   [Refereed]
Katsuhito Sudoh,Taro Watanabe,Jun Suzuki,Hajime Tsukada,Hideki Isozaki
2008 International Workshop on Spoken Language Translation, IWSLT 2008, Honolulu, Hawaii, USA, October 20-21, 2008   92-97   2008   [Refereed]
HIRAO Tsutomu, SUZUKI Jun, ISOZAKI Hideki
Transactions of the Japanese Society for Artificial Intelligence   22 574-584   Nov 2007
In the study of automatic summarization, the main research topic was `important sentence extraction' but nowadays `sentence compression' is a hot research topic. Conventional sentence compression methods usually transform a given sentence into a p...
Taro Watanabe,Jun Suzuki,Hajime Tsukada,Hideki Isozaki
EMNLP-CoNLL 2007, Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, June 28-30, 2007, Prague, Czech Republic   764-773   2007   [Refereed]
Jun Suzuki,Akinori Fujino,Hideki Isozaki
EMNLP-CoNLL 2007, Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning, June 28-30, 2007, Prague, Czech Republic   791-800   2007   [Refereed]
Taro Watanabe,Jun Suzuki,Katsuhito Sudoh,Hajime Tsukada,Hideki Isozaki
2007 International Workshop on Spoken Language Translation, IWSLT 2007, Trento, Italy, October 15-16, 2007   111-118   2007   [Refereed]
Jun Suzuki,Yutaka Sasaki,Eisaku Maeda
Systems and Computers in Japan   37(10) 58-68   2006   [Refereed]
Norihito Yasuda,Tsutomu Hirao,Jun Suzuki,Hideki Isozaki
Computational Approaches to Analyzing Weblogs, Papers from the 2006 AAAI Spring Symposium, Technical Report SS-06-03, Stanford, California, USA, March 27-29, 2006   231-236   2006   [Refereed]
Jun Suzuki,Erik McDermott,Hideki Isozaki
ACL 2006, 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Sydney, Australia, 17-21 July 2006      2006   [Refereed]
Taro Watanabe,Jun Suzuki,Hajime Tsukada,Hideki Isozaki
2006 International Workshop on Spoken Language Translation, IWSLT 2006, Keihanna Science City, Kyoto, Japan, November 27-28, 2006   95-102   2006   [Refereed]
HIRAO TSUTOMU, SUZUKI JUN, ISOZAKI HIDEKI, MAEDA EISAKU
IPSJ journal   46(10) 2533-2545   Oct 2005
Monolingual aligned corpora are valuable for natural language processing. In order to generate text, we can learn various kinds of knowledge from such corpora. For instance, summary sentences aligned with sentences from original documents are usef...
SUZUKI Jun, SASAKI Yutaka, MAEDA Eisaku
The IEICE transactions on information and systems Pt. 2   88(2) 230-240   Feb 2005
本論文では, テキスト処理に適したカーネル「階層非循環有向グラフカーネル」(HDAGカーネル)を提案する.テキスト処理のあらゆるタスクにおいて, テキストのもつ情報をどのように表現し, どのように演算にのせるかということは, 常に我々が直面する重要な課題である.これまで, 出現単語の頻度や系列などに着目したカーネル関数が提案されているが, テキストの表層的な情報しか扱っていない, 表現形式の有効性が十分に検証されていない, などの問題点があった.そこで, 階層非循環有向グラフ, すなわち,...
Taku Kudo,Jun Suzuki,Hideki Isozaki
ACL 2005, 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, 25-30 June 2005, University of Michigan, USA   189-196   2005   [Refereed]
Hajime Tsukada,Taro Watanabe,Jun Suzuki,Hideto Kazawa,Hideki Isozaki
2005 International Workshop on Spoken Language Translation, IWSLT 2005, Pittsburgh, PA, USA, October 24-25, 2005   112-117   2005   [Refereed]
Jun Suzuki,Hideki Isozaki
Advances in Neural Information Processing Systems 18 [Neural Information Processing Systems, NIPS 2005, December 5-8, 2005, Vancouver, British Columbia, Canada]   1321-1328   2005   [Refereed]
SASAKI YUTAKA, ISOZAKI HIDEKI, SUZUKI JUN, KOKURYOU KOJI, HIRAO TSUTOMU, KAZAWA HIDETO, MAEDA EISAKU
Transactions of Information Processing Society of Japan   45(2) 635-646   Feb 2004
This paper describes a Japanese Question-Answering (QA) System, SAIQA-II. These years, researchers have been attracted to the study of developing Open-Domain QA systems that find answers to a natural language question given by a user. Most of conv...
Jun Suzuki,Hideki Isozaki,Eisaku Maeda
Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics, 21-26 July, 2004, Barcelona, Spain.   119-126   2004   [Refereed]
Tsutomu Hirao,Jun Suzuki,Hideki Isozaki,Eisaku Maeda
COLING 2004, 20th International Conference on Computational Linguistics, Proceedings of the Conference, 23-27 August 2004, Geneva, Switzerland      2004   [Refereed]
SUZUKI JUN, SASAKI YUTAKA, MAEDA EISAKU
Transactions of Information Processing Society of Japan   44(11) 2839-2853   Nov 2003
Question type classification attempts to identify the intention of a given question. The approach to high performance question classification typically yields an extremely large number of features because question types are well featured by the st...
Jun Suzuki,Tsutomu Hirao,Yutaka Sasaki,Eisaku Maeda
Proceedings of the 41st Annual Meeting of the Association for Computational Linguistics, 7-12 July 2003, Sapporo Convention Center, Sapporo, Japan.   32-39   2003   [Refereed]
Jun Suzuki,Yutaka Sasaki,Eisaku Maeda
Advances in Neural Information Processing Systems 16 [Neural Information Processing Systems, NIPS 2003, December 8-13, 2003, Vancouver and Whistler, British Columbia, Canada]   643-650   2003   [Refereed]
Jun Suzuki,Yutaka Sasaki,Eisaku Maeda
19th International Conference on Computational Linguistics, COLING 2002, Howard International House and Academia Sinica, Taipei, Taiwan, August 24 - September 1, 2002      2002   [Refereed]

Misc

 
KAWAMURA Keigo, SUZUKI Jun, TSURUOKA Yoshimasa
Proceedings of the Annual Conference of JSAI   2018(0) 1N301-1N301   2018
<p>Neural fictitious self-play (NFSP) is a method for solving imperfect information games. While methods developed in recent years such as counterfactual regret minimization or DeepStack require the state transition rules of the games, NFSP works ...
Daisuke Taniuchi, Takuya Maekawa, Jun Suzuki, Yasue Kishino
研究報告ユビキタスコンピューティングシステム(UBI)   2013(5) 1-7   Oct 2013
Recently, many indoor positioning techniques based on Wi-Fi signals have been studied. Wi-Fi fingerprinting technique, which is one of the most popular and practical method, makes use of the Wi-Fi received signal strength (RSS) information collect...
Daisuke Taniuchi, Takuya Maekawa, Jun Suzuki, Yasue Kishino
IPSJ SIG technical reports   2013(5) 1-7   Oct 2013
Recently, many indoor positioning techniques based on Wi-Fi signals have been studied. Wi-Fi fingerprinting technique, which is one of the most popular and practical method, makes use of the Wi-Fi received signal strength (RSS) information collect...
数原 良彦, 鈴木 潤, 鷲崎 誠司
人工知能学会全国大会論文集   27 1-4   2013
須藤 克仁, 鈴木 潤, 塚田 元
Japio year book   292-296   2013
Suhara Yoshihiko, Suzuki Jun, Kataoka Ryoji
情報科学技術フォーラム講演論文集   11(2) 91-92   Sep 2012
SUZUKI Jun, FUJINO Akinori, ISOZAKI Hideki
IPSJ SIG Notes   2007(94) 21-28   Sep 2007
This paper proposes a novel semi-supervised conditional random field which provides good characteristics with respect to handling the large and sparse feature spaces. Experiments on two real NLP tasks with extremely large feature spaces, such as n...
FUJINO Akinori, ISOZAKI Hideki, SUZUKI Jun
IPSJ SIG Notes   2007(94) 29-34   Sep 2007
Text categorization is generally defined as a multi-labeling problem, where multiple category labels are assigned to each text document. We focus on machine learning approaches to multi-labeling problems and present a classifier design method base...
HIRAO Tsutomu, SUZUKI Jun, ISOZAKI Hideki, MAEDA Eisaku
IPSJ SIG Notes   2003(108) 31-38   Nov 2003
In this paper, we propose a multiple document summarization method using a sequential pattern mining algorithm. We extract important sentences in the following way ; First, extracting term patterns from target docment set by using PrefixSpan. Seco...
ISOZAKI Hideki, HIRAO Tsutomu, SUZUKI Jun
IPSJ SIG Notes   2003(108) 63-68   Nov 2003
Machine Learning is used for various tasks of Natural Language processing such as Named Entity Rcognition, Important Sentence Extraction, and Dependency Analysis. Features for Machine Learning are found by trial and error. However, it is possible ...
Hirao Tsutomu, Suzuki Jun, Isozaki Hideki, Maeda Eisaku
IEICE technical report. Natural language understanding and models of communication   103(407) 31-38   Nov 2003
In this paper, we propose a multiple document summarization method using a sequential pattern mining algorithm. We extract important sentences in the following way; First, extracting term patterns from target docment set by using PrefixS- pan. Sec...
ISOZAKI Hideki, HIRAO Tsutomu, SUZUKI Jun
IEICE technical report. Natural language understanding and models of communication   103(407) 63-68   Nov 2003
Machine Learning is used for various tasks of Natural Language Processing such as Named Entity Recognition, Important Sentence Extraction, and Dependency Analysis. Features for Machine Learning are found by trial and error. However, it is possible...
HIRAO Tsutomu, SUZUKI Jun, ISOZAKI Hideki, MAEDA Eisaku
電子情報通信学会技術研究報告. NLC, 言語理解とコミュニケーション   103(407) 31-38   Oct 2003
ISOZAKI Hideki, HIRAO Tsutomu, SUZUKI Jun
電子情報通信学会技術研究報告. NLC, 言語理解とコミュニケーション   103(407) 63-68   Oct 2003
鈴木 潤, 平尾 努, 磯崎 秀樹, 前田 英作
情報処理学会研究報告自然言語処理(NL)   2003(98) 41-48   Sep 2003
本稿では,String Kernelに対する素性選択手法について議論する.StringKernelを含むConvolution Kernelsの枠組では,全ての部分構造間の部分カーネルの総和を全体のカーネルの値と定義している.しかし,可能な全ての部分構造を素性として使用すると,素性空間の次元数が高くなり,データスパースネスの問題が起こることが実験的に示されている.このためString Kernelでは,カーネル計算に使用する部分記号列を部分記号列のサイズに応じて選択する枠組が導入されている...
SUZUKI Jun, HIRAO Tsutomu, ISOZAKI Hideki, MAEDA Eisaku
IPSJ SIG Notes   2003(98) 41-48   Sep 2003
This paper discusses feature selection methods about String Kernel. The kernel value of Convolution Kernels is defined as a sum of all sub-kernels between all sub-structures of input objects. However, it is known experimentally that the datasparse...
SUZUKI Jun, HIRAO Tsutomu, SASAKI Yutaka, MAEDA Eisaku
IPSJ SIG Notes   2003(23) 101-108   Mar 2003
This paper proposes an efficient calculation method of the similarity between two texts which reflects various structures inside the texts. We realize this similarity by expressing the structures inside a text by hierarchical directed acyclic grap...
Suzuki Jun, Sasaki Yutaka, Maeda Eisaku
情報技術レターズ   1 89-90   Sep 2002
SASAKI Y., ISOZAKI H., TAIRA H., HIRAO T., KAZAWA H., SUZUKI J., KOKURYO K., MAEDA E.
IPSJ SIG Notes   2001(86) 77-82   Sep 2001
Conventional TREC-style Question-Answering (QA) involves extracting passages (250 bytes or 50 bytes) that contain answers to a question. The most attractive feature of QA is that it can provide exact answers to the question, rather than a list of ...
Sasaki Y., Isozaki H., Taira H., Hirao T., Kazawa H., Suzuki J., Kokuryo K., Maeda E.
IPSJ SIG Notes   2001(86) 77-82   Sep 2001
Conventional TREC-style Question-Answering (QA) involves extracting passages (250 bytes or 50 bytes) that contain answers to a question. The most attractive feature of QA is that it can provide exact answers to the question, rather than a list of ...