A method for team intention inference
International Journal of Human-Computer Studies
Deception Detection through Automatic, Unobtrusive Analysis of Nonverbal Behavior
IEEE Intelligent Systems
AI & Society - Special Issue: Social intelligence design: a junction between engineering and social sciences
Hi-index | 0.00 |
We usually speculate partner's mental states by diverse nonverbal information. Agents need the ability for natural communication with people. In this paper, we focused on a lie as one of the typical behavior in which we often express our mental states unconsciously. The purpose of this study is to experimentally investigate the possibility of automatic lie detection in communication. We proposed an experimental setting in which participants could spontaneously decide whether or not to tell a lie. We then conducted an experiment to record participants' behavior in this setting. After that, we investigated, by discriminant analysis, that we could achieve 68% accuracy in classifying the utterances into lies and the rest without taking account of individual features by using the noverbal behavior data. We would detect participants' stresses when they told a lie. The suggestions in this paper are useful to an agent which pays attention to user's mental states.