Monday, September 25, 2006
11:00 AM
CSB 209
James F. Allen
U. Rochester
Collaborative Dialogue Agents
While plan- or agent- based models provide interesting formalisms for producing dialogue agents, most robust dialogue system use much simpler technology based on finite state machines and/or slot filling models. Over the past decade, we have been trying to develop dialogue-based models that not only provide a general theoretical account for dialogue behavior, but also lead to practical, robust, working systems. Most recently, these systems have not only been able to execute hand-coded tasks, but can also learn new tasks from the user through combined demonstration and dialogue. I will also show running examples of most recent system as they learn and execute new procedures that involve finding and using information on the world wide web.