Building Multilingual Lexical-Semantic Web by Sense Linking

2016 Fiscal Year Final Research Report

• Project/Area Number: 25280117
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: Library and information science/Humanistic social informatics
• Research Institution: Waseda University (2014-2016); Osaka University (2013)
• Principal Investigator: Hayashi Yoshihiko 早稲田大学, 理工学術院, 教授 (任期付) (80379156)
• Co-Investigator(Kenkyū-buntansha): 加藤 恒昭 東京大学, 総合文化研究科, 教授 (60334299) ; 岩根 久 大阪大学, 言語文化研究科(研究院), 教授 (50176559)
• Research Collaborator: LAO Wentao 大阪大学, 大学院学生 ; KANADA Kentaro 早稲田大学, 大学院学生
• Project Period (FY): 2013-04-01 – 2017-03-31
• Keywords: 情報資源の構築・管理 / 意味属性 / 意味的類似度 / セマンティックWeb / リンクデータ / 分散表現 / 機械学習 / 概念の想起関係

Outline of Final Research Achievements

We studied a methodology for linking lexical-semantic resources with different conceptual systems in different languages on the basis of Linked Data. The key element of the proposal is “linking by meaning”: the probabilistic correspondences of concept-defining words and the semantic similarity between gloss texts are utilized in a machine learning-based framework. Through empirical results, we confirmed that the fundamental applicability of the method, and mined future research issues. Moreover, in accordance with the progress of the research, we carried out researches on the computation of cross-lingual semantic sentential similarities and the utilization of distributed representation of word meanings.

Free Research Field

言語情報処理