UWTV Program: Sharing and Abstraction in Hierarchical Reinforcement Learning

Search:

Search for information by:

Note: Selected programs are available for streaming download per agreement with the original media source.

Sponsored by:

Sharing and Abstraction in Hierarchical Reinforcement Learning

	One of the most interesting challenges for machine learning is learning for sequential decision-making (also known as Reinforcement Learning or RL). Tom Dietterich of Oregon State University describes the approach of Hierarchical Reinforcement Learning using the MAXQ value function decomposion, where the programmer defines a hierarchy of tasks and subtasks, and the value function is decomposed hierarchically into value functions for each task and subtask.



	* Help?


Series Title:	CSE Colloquia - 2002
Subject(s):	Engineering and Computer Science
Speaker(s):	Tom Dietterich, Oregon State University
Related Link(s):	CSE web site
Production Date:	05/21/2002
Runtime:	00:59:14
Rating:	TV-G

Support for UWTV is provided by:

San Jose State @ UW
Watch 9/3 at 8 p.m.
Original Game: 11/16/96

Kansas State @ UW
Watch 9/6 at 7 p.m.
Original Game: 9/28/91

Email us with comments or questions or call 888-616-UWTV