Expressing social attitudes in virtual agents for social training games

摘要

The use of virtual agents in social coaching has increased rapidly in thelast decade. In order to train the user in different situations than can occurin real life, the virtual agent should be able to express different socialattitudes. In this paper, we propose a model of social attitudes that enables avirtual agent to reason on the appropriate social attitude to express duringthe interaction with a user given the course of the interaction, but also theemotions, mood and personality of the agent. Moreover, the model enables thevirtual agent to display its social attitude through its non-verbal behaviour.The proposed model has been developed in the context of job interviewsimulation. The methodology used to develop such a model combined a theoreticaland an empirical approach. Indeed, the model is based both on the literature inHuman and Social Sciences on social attitudes but also on the analysis of anaudiovisual corpus of job interviews and on post-hoc interviews with therecruiters on their expressed attitudes during the job interview.