Designing Social Dialogue Model for Human-Robot Interactions

Abstract

Social interactions between humans and robots increase in diverse contexts. The overall purpose of the study is to build the social dialogue model by examining the detailed structure and elements of conversation between humans and robots. To design the social dialogue model, we collected conversation data using our own developed system. We recruited 80 participants who were randomly assigned to complete conversations through the system. After collecting the conversation data, we are currently in the process of tagging them with social cues such as self-disclosure, self-disclosure elicitation, acknowledgement, shared experience, and topic management. With the tagged data, we will build the social dialogue model with a series of social cues. The implemented social dialogue model could be applied to diverse conversational agents such as chatbots.