Conversational agent who achieves tasks while interacting with humans based on scenarios

Abstract

Many interaction-oriented robots have been proposed as participants in human society. These agents, however, are not committed to achieving tasks. To become our partners, these agents should not only interact with humans but also achieve tasks with humans. In this paper, we propose a conversational agent who achieves tasks while interacting with humans based on scenarios. To realize this type of agent, we develop an enhanced BDI architecture by effectively integrating the research results of interaction-oriented and task-oriented agents. We show the effectiveness of the proposed agent by applying it to a conversational agent, called Photo-agent, who takes photographs while interacting with users.