Interpretation of Partial Utterances in Virtual Human Dialogue Systems

Kenji Sagae, David DeVault, David R. Traum

Abstract: Dialogue systems typically follow a rigid pace of interaction where the system waits until the user has finished speaking before producing a response. Interpreting user utterances before they are completed allows a system to display more sophisticated conversational behavior, such as rapid turn-taking and appropriate use of backchannels and interruptions. We demonstrate a natural language understanding approach for partial utterances, and its use in a virtual human dialogue system that can often complete a user's utterances in real time.