Exploration on context-sensitive affect sensing in an intelligent agent

Virtual drama systems generally promote users’ experience on learning/training. It is even challenging to produce intelligent conversational agents which are capable of engaging in drama performance and responding to human users’ emotion and language effectively and intelligently. In our previous work, we developed an intelligent agent embedded in a text-based improvisational virtual drama system to interact with human users and detect affect from their text input. The detection was solely based on the interpretation of each turn-taking input. In this paper, we further equip the AI agent with the capabilities of the inferences of the improvisational mood of individual characters and emotions embedded in communication context to further verify the affect drawn solely from the analysis of each turn-taking input itself. Evaluations results on context-based affect interpretation are also provided. Our work shows great potentials in employing emotions embedded in the characters and interaction context to provide personalized learning/training experience.