Construction of a computational model of word sense based on vocabulary, phonology, and pronunciation, and its application to multiple document summarization

2022 Fiscal Year Final Research Report

• Project/Area Number: 18K11429
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 61030:Intelligent informatics-related
• Research Institution: University of Yamanashi
• Principal Investigator: SUZUKI Yoshimi 山梨大学, 大学院総合研究部, 教授 (20206551)
• Project Period (FY): 2018-04-01 – 2023-03-31
• Keywords: 語感 / 感情 / 筆者の意図

Outline of Final Research Achievements

By using vocabulary, phonology, and pronunciation effectively, we were able to distinguish nuance in words. The use of word sense information enables us to estimate the author's intention and paves the way for the creation of summaries that emphasize the author's intention. Based on these findings, we developed a sentiment analysis of various review sentences, a peer-review score prediction system using peer-reviewed sentences, an article writing support system, and a speech synthesis model using facial landmarks, and our research results were published.

Free Research Field

自然言語処理

Academic Significance and Societal Importance of the Research Achievements

語彙と音韻及び発音を利用することで語感の区別ができるようになった．また，語感情報により筆者の意図を予測することが可能になり，より効率の良い要約への利用につなげることができた．