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

Text semantic parsing combining dynamic knowledge obtained from preceding context and static knowledge obtained in pre-training

Research Project

Project/Area Number 19K12112
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 InstitutionTohoku University

Principal Investigator

Matsubayashi Yuichiroh  東北大学, 教育学研究科, 准教授 (20582901)

Project Period (FY) 2019-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2021: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2020: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2019: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
Keywords省略解析 / 述語項構造 / 意味解析 / 文章理解 / 自然言語処理 / 意味構造解析 / 人工知能
Outline of Research at the Start

文を「正確に読む」ために必須となる省略解析の技術の性能は、未だ一般的な人間の読解能力と比べて大きな隔たりがある。本研究課題では、文章の一部(多くは1、2文単位)だけを見て解析を行う従来型の意味解析手法を改良し、正確な意味理解に不可欠な (1) 先行文脈の文意の蓄積(=動的知識)に基づいて後方の意味を理解するモデルと (2) 推論に必要な常識的知識(=静的知識)の効果的・効率的な表現方法の確立、(3) これら動的知識と静的知識を融合した自然な推論に基づく意味解析の実現を行うことで意味解析技術の大幅な向上を目指す。

Outline of Final Research Achievements

The performance of omission analysis, which is fundamental to "accurate reading" of sentences, had a large gap compared to general human reading comprehension ability. The goal of this research was to improve the conventional semantic parsing technique, which only looks at a few sentences around the target sentence, by (1) constructing a parsing model that understands meaning based on accumulating the meaning of sentences in the preceding context (= dynamic knowledge), (2) establishing an effective and efficient method to express the common sense knowledge required for inference (= static knowledge), and (3) realizing the natural inference based on the combination of such dynamic and static knowledge. As a result, we were successful in implementing an omission analysis system that significantly outperformed previous methods' performance. Our proposed system, developed in this project, has been publicly released as open source software.

Academic Significance and Societal Importance of the Research Achievements

省略解析は文章の意味を正確に理解するAIの実現に不可欠な要素であり、日本語解析のボトルネックとなっていたこの基盤技術の解析精度向上により、応用技術の発展可能性が増大した意義は大きい。開発したシステムは一般公開し、実世界テキスト解析に適用可能である。精度向上の鍵となったアイデアは、汎用的言語モデルに対して学習の形態を大きく変更することなくシームレスに意味解析の能力を増強するものであり、その他の言語処理技術の性能向上に対しても応用可能性を秘めている。加えて、研究過程で得られた知見から書き手の省略判断分析という新たな研究の方向性を展開し、教育応用等へのシードを生んだ点も学術的意義として挙げられる。

Report

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

    (17 results)

All 2023 2022 2021 2020

All Journal Article (6 results) (of which Peer Reviewed: 6 results,  Open Access: 6 results) Presentation (11 results)

  • [Journal Article] Reducing the Cost: Cross-Prompt Pre-Finetuning for Short Answer Scoring2023

    • Author(s)
      Hiroaki Funayama, Yuya Asazuma, Yuichiroh Matsubayashi, Tomoya Mizumoto and Kentaro Inui
    • Journal Title

      Artificial Intelligence in Education. AIED 2023.

      Volume: --

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Balancing Cost and?Quality: An Exploration of?Human-in-the-Loop Frameworks for?Automated Short Answer Scoring2022

    • Author(s)
      Funayama Hiroaki、Sato Tasuku、Matsubayashi Yuichiroh、Mizumoto Tomoya、Suzuki Jun、Inui Kentaro
    • Journal Title

      Artificial Intelligence in Education. AIED 2022. Lecture Notes in Computer Science

      Volume: 13355 Pages: 465-476

    • DOI

      10.1007/978-3-031-11644-5_38

    • ISBN
      9783031116438, 9783031116445
    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Pseudo Zero Pronoun Resolution Improves Zero Anaphora Resolution2021

    • Author(s)
      Konno Ryuto、Kiyono Shun、Matsubayashi Yuichiroh、Ouchi Hiroki、Inui Kentaro
    • Journal Title

      Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing

      Volume: NA Pages: 3790-3806

    • DOI

      10.18653/v1/2021.emnlp-main.308

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Multi-dialect Neural Machine Translation for 48 Low-resource Japanese Dialects2020

    • Author(s)
      Abe Kaori、Matsubayashi Yuichiroh、Okazaki Naoaki、Inui Kentaro
    • Journal Title

      Journal of Natural Language Processing

      Volume: 27 Issue: 4 Pages: 781-800

    • DOI

      10.5715/jnlp.27.781

    • NAID

      130007998363

    • ISSN
      1340-7619, 2185-8314
    • Year and Date
      2020-12-15
    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] An Empirical Study of Contextual Data Augmentation for Japanese Zero Anaphora Resolution2020

    • Author(s)
      Konno Ryuto、Matsubayashi Yuichiroh、Kiyono Shun、Ouchi Hiroki、Takahashi Ryo、Inui Kentaro
    • Journal Title

      Proceedings of the 28th International Conference on Computational Linguistics

      Volume: 1 Pages: 4956-4968

    • DOI

      10.18653/v1/2020.coling-main.435

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Preventing Critical Scoring Errors in Short Answer Scoring with Confidence Estimation2020

    • Author(s)
      Funayama Hiroaki、Sasaki Shota、Matsubayashi Yuichiroh、Mizumoto Tomoya、Suzuki Jun、Mita Masato、Inui Kentaro
    • Journal Title

      Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop

      Volume: 1 Pages: 237-243

    • DOI

      10.18653/v1/2020.acl-srw.32

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access
  • [Presentation] 日本語話者の項省略判断に関するアノテーションとモデリング2023

    • Author(s)
      石月由紀子, 栗林樹生, 松林優一郎, 笹野遼平, 乾健太郎
    • Organizer
      言語処理学会第29回年次大会
    • Related Report
      2022 Annual Research Report
  • [Presentation] 文章構造グラフを用いた国語記述式答案への自動フィードバック生成2023

    • Author(s)
      岩瀬裕哉, 舟山弘晃, 松林優一郎, 乾健太郎
    • Organizer
      言語処理学会第29回年次大会
    • Related Report
      2022 Annual Research Report
  • [Presentation] What can Short Answer Scoring Models Learn from Cross-prompt Training Data?2023

    • Author(s)
      Hiroaki Funayama, Yuya Asazuma, Yuichiroh Matsubayashi, Tomoya Mizumoto, Kentaro Inui.
    • Organizer
      言語処理学会第29回年次大会
    • Related Report
      2022 Annual Research Report
  • [Presentation] 情報量に基づく日本語項省略の分析2022

    • Author(s)
      石月由紀子, 栗林樹生, 松林優一郎, 大関洋平
    • Organizer
      言語処理学会第28回年次大会
    • Related Report
      2021 Research-status Report
  • [Presentation] 記述式答案自動採点における確信度推定とその役割2022

    • Author(s)
      舟山弘晃, 佐藤汰亮, 松林優一郎, 水本智也, 鈴木潤, 乾健太郎
    • Organizer
      言語処理学会第28回年次大会
    • Related Report
      2021 Research-status Report
  • [Presentation] 事前学習とfinetuningの類似性に基づくゼロ照応解析2021

    • Author(s)
      今野颯人, 松林優一郎, 清野舜, 大内啓樹, 乾健太郎
    • Organizer
      言語処理学会第27回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] 項目採点技術に基づいた和文英訳答案の自動採点2021

    • Author(s)
      菊地正弥, 尾中大介, 舟山弘晃, 松林優一郎, 乾健太郎
    • Organizer
      言語処理学会第27回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] 実用的な自動採点のための確信度推定と根拠事例の提供2021

    • Author(s)
      舟山弘晃,王天奇,松林優一郎,水本智也,佐藤汰亮,鈴木潤,乾健太郎
    • Organizer
      言語処理学会第27回年次大会
    • Related Report
      2020 Research-status Report
  • [Presentation] マスク言語モデルを利用したデータ拡張に基づく日本語文内ゼロ照応解析2020

    • Author(s)
      今野颯人, 松林優一郎, 清野舜, 大内啓樹, 高橋諒, 乾健太郎
    • Organizer
      言語処理学会第26回年次大会
    • Related Report
      2019 Research-status Report
  • [Presentation] 反復改良法を用いた日本語述語項構造解析2020

    • Author(s)
      宮脇峻平, 清野舜, 松林優一郎, 今野颯人, 高橋諒, 大内啓樹, 乾健太郎
    • Organizer
      言語処理学会第26回年次大会
    • Related Report
      2019 Research-status Report
  • [Presentation] 記述式答案自動採点のための確信度推定手法の検討2020

    • Author(s)
      舟山弘晃, 佐々木翔大, 水本智也, 三田雅人, 鈴木潤, 松林優一郎, 乾健太郎
    • Organizer
      言語処理学会第26回年次大会
    • Related Report
      2019 Research-status Report

URL: 

Published: 2019-04-18   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi