Development of an interactive agent that acquires user preferences from affective information

2023 Fiscal Year Final Research Report

• Project/Area Number: 20K19905
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 61060:Kansei informatics-related
• Research Institution: Nagoya Institute of Technology
• Principal Investigator: Sakuma Takuto 名古屋工業大学, 工学(系)研究科(研究院), 助教 (20753627)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: 感情推定 / 感情表現 / ヒューマンエージェントインタラクション

Outline of Final Research Achievements

In this research, the development of technologies for emotion estimation and emotion expression has been promoted, and the results have been reported in journals and international conferences. In particular, the performance of models that estimate emotions from speech, facial expressions, and brain waves has been improved, and technology has been developed for agents that interact with users in a friendly manner, mainly in text chats. In addition, research results on a dementia diagnosis system based on speech sounds using speech information processing technology have accumulated up to now, and results on developing a BCI system utilizing the knowledge of cerebral blood flow analysis have also been presented.

Free Research Field

感性情報処理

Academic Significance and Societal Importance of the Research Achievements

感情推定技術を用いることで，ユーザの嗜好を獲得し，インタラクションにユーザの好みを反映することが出来るエージェントの構築が本研究課題の最終目標であり，このようなエージェントはこれまで学術的にも実現出来ていない．昨今，人工知能が身近な存在となりつつある中で人に寄り添うエージェントの実現は社会的にも意義が大きい．また，研究成果の概要でも述べた認知症診断システムおよびBCIシステム開発の社会的意義の大きさは明らかである．