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

2020 Fiscal Year Final Research Report

Development of a multi-purpose categorial grammar treebank

Research Project

  • PDF
Project/Area Number 18K00523
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 02060:Linguistics-related
Research InstitutionNational Institute for Japanese Language and Linguistics (2019-2020)
University of Tsukuba (2018)

Principal Investigator

Kubota Yusuke  大学共同利用機関法人人間文化研究機構国立国語研究所, 理論・対照研究領域, 准教授 (60745149)

Co-Investigator(Kenkyū-buntansha) 峯島 宏次  慶應義塾大学, 文学部, 准教授 (80725739)
Project Period (FY) 2018-04-01 – 2021-03-31
Keywordsツリーバンク / カテゴリ文法 / 意味解析 / 日本語 / NPCMJ / ABCツリーバンク
Outline of Final Research Achievements

In this project, we constructed a categorial grammar treebank called the ABC Treebank. The ABC Treebank has been made available as an open-access linguistic resource for interdisciplinary research in theoretical linguistics and natural language processing. A characteristic feature of this treebank is linguistically adequate analyses of sentence-final predicates such as causatives and passives as well as the structure of noun modifications. We have released related tools such as the script used for treebank conversion and the depccg parser training model using the treebank. In addition, a pilot experiment was conducted to derive semantic representations using the depccg parser and the ccg2lambda toolkit, and it was confirmed that the treebank can be used as a framework for semantic analysis in NLP research.

Free Research Field

言語学(統語論、意味論)

Academic Significance and Societal Importance of the Research Achievements

ABCツリーバンクは、日本語の範疇文法ツリーバンクで公開されているものとして現在唯一のものであり、理論言語学、自然言語処理両分野を取り結ぶ学際的な研究領域の活性化に寄与することが期待される。特に、パーザやツリーバンクなど、自然言語処理のツールを援用した計算論的モデリングに基づく言語研究が近年理論言語学研究において活発になってきているが、計算言語学の究極の課題である、意味解析・意味理解にまで踏み込んだ研究はまだ極めて少ない。範疇文法に基づくABCツリーバンクは、意味解析・意味理解の計算論的研究のための基盤資源としての利用が特に期待できる。

URL: 

Published: 2022-01-27  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi