• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Word Sense Disambiguation Using Semi-supervised Deep Learning

Research Project

Project/Area Number 18K11422
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 61030:Intelligent informatics-related
Research InstitutionIbaraki University

Principal Investigator

Sasaki Minoru  茨城大学, 理工学研究科(工学野), 准教授 (60344834)

Project Period (FY) 2018-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥3,510,000 (Direct Cost: ¥2,700,000、Indirect Cost: ¥810,000)
Fiscal Year 2020: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2019: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2018: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Keywords語義曖昧性解消 / 機械学習 / グラフニューラルネットワーク / 半教師あり学習 / 自然言語処理 / グラフベース手法 / 意味解析 / 半教師あり深層学習
Outline of Final Research Achievements

In this study, we developed a semi-supervised WSD method using semantic similarities between example sentences. In this method, we propose a graph construction method that does not require any parameters using BERT pre-trained model to represent a semantic similarity relation obtained from sense labeled examples and unlabeled examples. As a result of evaluating the effectiveness of the system, the developed system improved the accuracy of word sense identification by 1.73% compared to an existing Japanese semi-supervised word sense disambiguation system. In addition, the results of a word sense disambiguation experiment using the SENSEVAL-2 English Lexical Task data, which is English assessment data, showed a 3% improvement in accuracy compared to the previous method, which achieved the highest accuracy. These results show that the developed system is effective in semi-supervised word sense disambiguation.

Academic Significance and Societal Importance of the Research Achievements

語義曖昧性解消において、「語義曖昧性解消をシンプルな半教師ありディープラーニングを使ったモデルで構築できないか」「少量の語義付き用例文を利用して語義の特徴を捉えたディープラーニングモデルを構築できないか」という2つの課題を解決する効果的な手法を確立することができた。
本研究の成果から得られる学術的な意義は、語義付き例文が少量のみ存在する場合でも従来手法では捉えられなかった効果的な文脈情報の取得や用例文間の意味的な関係の取得が可能となったことである。この成果により、用例文を大量に追加して効果的な識別モデルの学習が可能なことや用例文を大量に提供可能な国語辞典の編纂が可能となるなどの社会的意義がある。

Report

(6 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • 2019 Research-status Report
  • 2018 Research-status Report
  • Research Products

    (26 results)

All 2023 2021 2020 2019 2018

All Journal Article (3 results) (of which Peer Reviewed: 3 results,  Open Access: 2 results) Presentation (23 results) (of which Int'l Joint Research: 11 results,  Invited: 1 results)

  • [Journal Article] 用例文間の意味的な類似関係を用いた半教師あり語義曖昧性解消2021

    • Author(s)
      谷田部梨恵, 佐々木稔
    • Journal Title

      情報処理学会論文誌

      Volume: 62 Pages: 1724-1736

    • NAID

      170000185635

    • Related Report
      2021 Research-status Report
    • Peer Reviewed
  • [Journal Article] Unsupervised All-words WSD Using Synonyms and Embeddings2019

    • Author(s)
      Suzuki Rui、Komiya Kanako、Asahara Masayuki、Sasaki Minoru、Shinnou Hiroyuki
    • Journal Title

      Journal of Natural Language Processing

      Volume: 26 Issue: 2 Pages: 361-379

    • DOI

      10.5715/jnlp.26.361

    • NAID

      130007706831

    • ISSN
      1340-7619, 2185-8314
    • Year and Date
      2019-06-15
    • Related Report
      2019 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Comparison of Methods to Annotate Named Entity Corpora2018

    • Author(s)
      Kanako Komiya, Masaya Suzuki, Tomoya Iwakura, Minoru Sasaki, Hiroyuki Shinnou
    • Journal Title

      Transactions on Asian and Low-Resource Language Information Processing

      Volume: 34 Issue: 4 Pages: 1-16

    • DOI

      10.1145/3218820

    • Related Report
      2018 Research-status Report
    • Peer Reviewed / Open Access
  • [Presentation] WordNet Lexicographerカテゴリ推定による語義サイズ縮約を用いた語義曖昧性解消2023

    • Author(s)
      橋口卓弥、佐々木稔
    • Organizer
      言語処理学会第29回年次大会
    • Related Report
      2022 Annual Research Report
  • [Presentation] Japanese Word Sense Disambiguation Using Gloss Information of a Japanese Dictionary2021

    • Author(s)
      Hiroki Okemoto, Minoru Sasaki
    • Organizer
      the Thirteenth International Conference on Information, Process, and Knowledge Management (eKnow2021)
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] Person Name Extraction from TV program Using Pre-trained Language Model and News Headlines2021

    • Author(s)
      Kazuki Oda, Minoru Sasaki
    • Organizer
      the 12th International Conference on E-Service and Knowledge Management (ESKM 2021)
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] The reliability of word meanings in online dictionaries and how word meanings change over time2021

    • Author(s)
      Minoru Sasaki
    • Organizer
      The Thirteenth International Conference on Pervasive Patterns and Applications (PATTERNS2021)
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research / Invited
  • [Presentation] 辞書の階層構造埋め込み学習における日本語辞書定義文の効果的な利用2021

    • Author(s)
      石井佑樹, 佐々木稔
    • Organizer
      言語処理学会第26回年次大会
    • Related Report
      2021 Research-status Report
  • [Presentation] 語義の例文を使用した語義曖昧性解消の有効性分析2021

    • Author(s)
      関谷洸, 佐々木稔
    • Organizer
      言語処理学会第26回年次大会
    • Related Report
      2021 Research-status Report
  • [Presentation] 訓練事例と辞書用例を異なるモデルで表現した語義曖昧性解消2021

    • Author(s)
      谷田部梨恵, 佐々木稔
    • Organizer
      言語処理学会第27回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] Semi-supervised Word Sense Disambiguation Using Example Similarity Graph2020

    • Author(s)
      Rie Yatabe, Minoru Sasaki
    • Organizer
      Proceedings of the 14th Workshop on Graph-Based Natural Language Processing (TextGraphs-14)
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] Word Sense Disambiguation Using Graph-based Semi-supervised Learning2020

    • Author(s)
      Rie Yatabe, Minoru Sasaki
    • Organizer
      Proceedings of The Fourteenth International Conference on Advances in Semantic Processing (SEMAPRO2020)
    • Related Report
      2020 Research-status Report
  • [Presentation] 語義曖昧性解消における辞書に定義された単義語利用についての分析2020

    • Author(s)
      佐々木稔, 谷田部梨恵
    • Organizer
      言語資源活用ワークショップ2020
    • Related Report
      2020 Research-status Report
  • [Presentation] BERTの学習済みモデルを用いた用例文ペアの同義判定2020

    • Author(s)
      谷田部梨恵, 佐々木稔
    • Organizer
      言語処理学会第26回年次大会
    • Related Report
      2019 Research-status Report
  • [Presentation] NTCIR-15 QA Lab-PoliInfo2 のタスク設計2020

    • Author(s)
      木村泰知, 渋木英潔, 高丸圭一 , 秋葉友良, 石下円香, 内田ゆず, 小川泰弘, 乙武北斗, 佐々木稔, 三田村照子, 横手健一, 吉岡真治, 神門典子
    • Organizer
      言語処理学会第26回年次大会
    • Related Report
      2019 Research-status Report
  • [Presentation] Active Learning to Select Unlabeled Examples with Effective Features for Document Classification2019

    • Author(s)
      Minoru Sasaki
    • Organizer
      The 10th International Conference on Computational Linguistics and Intelligent Text Processing
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] Ibrk at the NTCIR-14 QA Lab-PoliInfo Classification Task2019

    • Author(s)
      Minoru Sasaki, Tetsuya Nogami
    • Organizer
      The Fourteenth NTCIR conference (NTCIR-14)
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] BERTモデルとニュースヘッドラインによるAI運用システムの試作2019

    • Author(s)
      史文愷, 細木唯以, 三好勝博, 江口潤一, 佐々木稔, 鈴木智也
    • Organizer
      日本機械学会2019年茨城講演会
    • Related Report
      2019 Research-status Report
  • [Presentation] グラフニューラルネットワークを用いた半教師あり語義曖昧性解消2019

    • Author(s)
      谷田部梨恵, 佐々木稔
    • Organizer
      情報処理学会 第241回自然言語処理研究会
    • Related Report
      2019 Research-status Report
  • [Presentation] 半教師あり語義曖昧性解消における各ジャンルの語義なし用例文の利用2019

    • Author(s)
      谷田部梨恵, 佐々木稔
    • Organizer
      言語資源活用ワークショップ2019
    • Related Report
      2019 Research-status Report
  • [Presentation] 単語区切りの違いによるQAサイトの質問回答ペアの分類2019

    • Author(s)
      佐々木稔, 古宮嘉那子
    • Organizer
      IDRユーザフォーラム2019
    • Related Report
      2019 Research-status Report
  • [Presentation] All-words Word Sense Disambiguation Using Concept Embeddings2018

    • Author(s)
      Rui Suzuki, Kanako Komiya, Masayuki Asahara, Minoru Sasaki, Hiroyuki Shinnou
    • Organizer
      Proceedings of the 11th edition of the Language Resources and Evaluation Conference
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] Detecting Unknown Word Senses in Contemporary Japanese Dictionary from Corpus of Historical Japanese2018

    • Author(s)
      Aya Tanabe, Kanako Komiya, Masayuki Asahara, Minoru Sasaki, Hiroyuki Shinnou
    • Organizer
      The 8th Conference of Japanese Association for Digital Humanities
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] Multi-Domain Word Embeddings for Semantic Relation Analysis among Domains2018

    • Author(s)
      Minoru Sasaki
    • Organizer
      Proceedings of The Fourth Asia Pacific Corpus Linguistics Conference
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] Word Embeddings of Monosemous Words in Dictionary for Word Sense Disambiguation2018

    • Author(s)
      Minoru Sasaki
    • Organizer
      Proceedings of The Twelfth International Conference on Advances in Semantic Processing
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research
  • [Presentation] Fine-tuning for Named Entity Recognition Using Part-of-Speech Tagging2018

    • Author(s)
      Masaya Suzuki, Kanako Komiya, Minoru Sasaki and Hiroyuki Shinnou
    • Organizer
      The 32th Pacific Asia Conference on Language, Information and Computation
    • Related Report
      2018 Research-status Report
    • Int'l Joint Research

URL: 

Published: 2018-04-23   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi