Conversation as Action Under Uncertainty

Proceedings of the Sixteenth Conference on Uncertainty in Artificial Intelligence |

Conversations abound with uncertainties of various kinds. Treating conversation as inference and decision making under uncertainty, we propose a task independent, multimodal architecture for supporting robust continuous spoken dialog called Quartet. We introduce four interdependent levels of analysis, and describe representations, inference procedures, and decision strategies for managing uncertainties within and between the levels. We highlight the approach by reviewing interactions between a user and two spoken dialog systems developed using the Quartet architecture: Presenter, a prototype system for navigating Microsoft PowerPoint presentations, and the Bayesian Receptionist, a prototype system for dealing with tasks typically handled by front desk receptionists at the Microsoft corporate campus.