人間-エージェントの相互適応系における継続的相互作用実現のための機能条件

概要

論文の詳細を見る
In the field of Robotics and AI, there have been gradually increasing the works dealing with Human-Agent Interactions, e.g. pet/humanoid robots developments/applications, interactive teaching, affective computing and so on. However, most of these studies aimed at making robots intelligent, and/or the developed agents mostly depended on human's unavoidable nature, that is, we can implicitly adapt to artificial agents. In contrast, it recently becomes clarified that mutual interaction/adaptation is the critical factor for the interaction design between human and artificial agents. This framework is referred to as HAI (Human-Agent Interaction) design. In this study, we especially focused on a crucial problem in HAI design framework, that is, how to keep a continuing interaction between a human and an agent/robot, because human could be apt to lose interest for agent as time goes on. Due to actualize continuing interactions between agents and the users, it seems to be crucial that both systems possess intrinsic/epigenetic adaptation abilities, and the agent should have an evaluation criteria for estimating internal conditions of users (e.g. curiosity). Based on this consideration, we discuss key functions for realizing a continuing HAI which the agent should possess.
2003-10-08