Host: The Japanese Society for Artificial Intelligence
Name : The 33rd Annual Conference of the Japanese Society for Artificial Intelligence, 2019
Number : 33
Location : [in Japanese]
Date : June 04, 2019 - June 07, 2019
For natural conversation with people, dialogue system has to return an appropriate reaction depending on dialogue mood.In this study, we aim to develop a method of estimating dialogue mood for that.We conducted the experiment to collect dialogue speech data, dialogue mood evaluation, and personality traits of subjects.Dialogue speech data and dialogue mood evaluation was used for calculating intervals of utterance state. Our method estimated dialogue mood by these feature values.In this paper, we verify effectiveness of perfonality traits added this time and the method of dialogue mood estimation.As a result, personality traits are effective on dialogue mood estimation in a part of dialogue mood labels, and the result suggests relevance of personality features with dialogue mood.