2018 Fiscal Year Final Research Report

Cognitive Analysis and Modeling of Human-Human Interaction

Planned Research

PDF

Project Area	Cognitive Interaction Design: A Model-Based Understanding of Communication and its Application to Artifact Design
Project/Area Number	26118002
Research Category	Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area)
Allocation Type	Single-year Grants
Review Section	Complex systems
Research Institution	The University of Tokyo
Principal Investigator	UEDA Kazuhiro 東京大学, 大学院総合文化研究科, 教授 (60262101)
Co-Investigator(Kenkyū-buntansha)	竹内勇剛静岡大学, 情報学部, 教授 (00333500) 峯松信明東京大学, 大学院工学系研究科(工学部), 教授 (90273333) 大本義正京都大学, 情報学研究科, 助教 (90511775) 本田秀仁安田女子大学, 心理学部, 講師 (60452017) 遠山紗矢香静岡大学, 情報学部, 助教 (80749664)
Research Collaborator	SAKAMOTO Takafumi YOSHIOKA Genta
Project Period (FY)	2014-07-10 – 2019-03-31
Keywords	認知科学 / 人工知能 / ヒューマンインタフェース / コミュニケーション / 社会系心理学
Outline of Final Research Achievements	First, focusing on the interaction that affects emotional state to the other party, we modeled spatial approach behavior based on internal states between oneself and others at the beginning of interaction. It was shown that the internal states of others could be estimated from their behavior. Second, we analyzed the dialogues and various kinds of non-verbal information when a clerk of the travel agency found out a customer's request and suggested travel plans. We could clarify the influence of temporal transition of verbal and non-verbal information on the success or failure of interaction between a clerk and a customer. Finally, we analyzed miscommunication caused by thick accents of non-native language speakers. The shadowing speech of a native language listener was used to quantify how thick the accents of a non-native speaker were, and we could objectively predict, from the behavior of native shadowers, subjective "listening difficulty" caused by thick accents.
Free Research Field	認知科学，知能情報学
Academic Significance and Societal Importance of the Research Achievements	人同士のインタラクションは，コミュニケーション欲求に基づく関係形成フェーズ，関係形成後の他者の内的状態の推定に基づく他者理解フェーズ，さらに理解の齟齬の検出に基づく相互調整フェーズの3つのフェーズからなる。研究1で最初のフェーズを，研究2-1と2-2で残りの2フェーズを分析でき，かつインタラクション相手の内部状態を推定するための他者モデルを構築できた。これはインタラクションのモデル研究に大きな前進をもたらし得る。さらに，研究2-1の成果はビジネス交渉の教育に応用できる。また，研究2-2の成果は，非母語話者に対する効果的な言語教育の実現という社会的な需要に沿うものであり，社会的な意義は大きい。