Development and implementation of an organizationally collaborative dialogue system that organically responds to latent and manifest demands

2023 Fiscal Year Final Research Report

• Project/Area Number: 19K14317
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 09070:Educational technology-related
• Research Institution: The University of Tokushima
• Principal Investigator: SEKI Yosuke 徳島大学, 高等教育研究センター, 准教授 (10623704)
• Project Period (FY): 2019-04-01 – 2024-03-31
• Keywords: 対話システム / 情報推薦 / ユーモア生成 / システム開発 / 機械学習

Outline of Final Research Achievements

A system for disseminating information in response to users' latent and overt requests for university public relations activities is designed and developed.As a result of introducing the system developed at Tokushima University, the system was able to resolve constraints caused by uncertain intentions and the ability to collect information, and was able to support users' information gathering.In addition, in order to form a friendly relationship between the user and the dialogue system aiming sustainable dialogue, a method for the response character to express emotions and actions according to the content of the user's utterances is proposed, and methods generateing pseudo mishearing, nori-tsukkomi sentences, tatoe-tsukkomi sentences are proposed for humorous expressions are proposed.As a result of introducing a dialogue system that implemented some of the methods into Tokushima University, the methods was able to improve in familiarity with the system and ease of asking questions.

Free Research Field

知能情報処理

Academic Significance and Societal Importance of the Research Achievements

提案した推薦型対話システムは，例えば大学への進学希望者等，ユーザが把握すべき情報が不明な状況や漠然と興味がある情報を効率的に収集する場面で有効である．また，親和な関係を形成するための手法を雑談対話システムに実装することで，ユーモアを交えた雑談対話が実現されて，高齢者や入院患者等の対話希望者との親和的なコミュニケーション，さらには笑いによる血糖値の上昇抑制やストレス低下等の健康促進・医療効果が期待できる．