Machine Learning by Discovering Macro-operators
Glenn Iba
Thu Jan 10, 17, 24, 31, 07:30-09:00pm, 24-619
No enrollment limit, no advance sign up
Participants welcome at individual sessions (series)
Prereq: none
Macro-operators (compiled sequences of lower-level operators) are a powerful abstraction technique, enabling complex planning and problem-solving by both humans and computers. We will explore methods for discovering macro-operators in the context of puzzles and video games (e.g. Rubik's Cube and Tetris) and work to build these techniques into machine learning programs. This activity could lead to on-going UROP projects.
Web: http://web.mit.edu/esg/www/iap
Contact: Glenn Iba, 24-612, x3-7786, giba@alum.mit.edu
Sponsor: Experimental Study Group
Latest update: 14-Jan-2002
|
|
