A chat information service system using a humanoid robot

In this paper, we describe development of an information service system using a humanoid robot. This system provides information such as topical news articles through conversations. Our goal is to evolve a conversational robot that can carry on conversations and enlighten the user. We think that achieving fun conversations is the basis for creating real symbiotic robots. This robot provides news information whether users explicitly ask for it or not during the conversation in order to attract users to conversations with the robot. The developed conversation mechanism is a natural extension of chatbot. Priorities for possible answers in response to user inputs are determined according to the dialogue strategy. This scheme is used for searching items from a news article database. With this method, one does not have to define an explicit domain model. This system can use pre-fixed databases but also renewable ones. We demonstrate the proposed idea on the humanoid robot, ASIMO.