Clarifying Questions for Conversational Agents

Many conversational agents blend the functionalities of question-answering and chit-chat into a single system. This lays the foundation for a more interactive type of question-answering models which is the primary focus of the research project. Interaction between a user and the agent allows the user to ask an underdefined or ambiguous question, and, instead of the system returning its best response, the system is able to ask clarifying questions to help it hone in on an answer it can be more confident in. Similarly, when the user doesn’t like or agree with the agent’s answer, the system is able to elicit more information from the user that can help the system produce a better answer from the user’s perspective.

Faculty Supervisor:

Frank Rudzicz

Student:

Partner:

Microsoft Canada

Discipline:

Computer science

Sector:

Information and cultural industries; Professional, scientific and technical services

University:

University of Toronto

Program: