Dialogue continuation by utterance induction using multi-attribute sensitivity empathy

2017 Fiscal Year Final Research Report

• Project/Area Number: 15H02770
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Kansei informatics
• Research Institution: The University of Tokushima
• Principal Investigator: Fuketa Masao 徳島大学, 大学院社会産業理工学研究部(理工学域), 教授 (10304552)
• Co-Investigator(Kenkyū-buntansha): 任 福継 徳島大学, 大学院社会産業理工学研究部(理工学域), 教授 (20264947) ; 森田 和宏 徳島大学, 大学院社会産業理工学研究部(理工学域), 准教授 (20325252) ; 青江 順一 徳島大学, ソシオテクノサイエンス研究部, 教授 (90108853)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 感性情報学 / 対話処理

Outline of Final Research Achievements

The purpose of this research is to continue conversation by generating computer responses to unexpected utterance. For generating computer responses, multi-attribute sensitivities from questions of the computer and human's response are extracted, and empathic sensibilities are determined. Topic sensitivity, tense, and type of question are extracted from the question sentences, and topic sensitivity, emotion sensitivity, tense, affirmation / negation are extracted from the response sentence. The reply sentence was generated from these combinations. From experimental results, the accuracy of the response was 90.0%.

Free Research Field

自然言語処理