Interactive Task Learning
Abstract
This research project will expand and extend our previous research on interactive task learning, where an artificial autonomous agent learns multiple new tasks from natural interactions with a human instructor. Interactive task learning has the potential to allow agent users, instead of developers, to extend the tasks that an agent can perform, without programming. Our approach builds on prior research on cognitive architecture that provides the necessary representation, processing, and learning mechanisms. Our approach emphasizes mixed initiative interaction, where the human provides advice and information, and the agent actively asks questions to acquire the knowledge it needs. Moreover, the agent learns by being situated in the task with the instructor, and it attempts to perform the task as it gains knowledge. We will focus on how learning can improve interaction, how interaction and reasoning can improve learning, and how reasoning can help ensure the correctness of what is learned. These will greatly expand the applicability of interactive task learning, eliminating key hurdles in its broader adoption and use.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 30, 2018
- Source ID
- FA95501810168
Entities
People
- John E. Laird
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Michigan