Department of Computer Science
Wednesday 5th May 1999 (6 pm)
ABSTRACT
Reinforcement learning is the study of how an agent may learn to perform a task, given feedback in the form of rewards or punishments. The task may have many stages, and rewards may be delayed so that unpleasant preparatory actions may be necessary to obtain rewards later. Originally inspired by analyses of animal learning, reinforcement learning is starting to be applied to control, scheduling, and routing systems.
This talk will describe how reinforcement learning problems can be formalised, and will go on to survey recent demonstrations of learning applied to realistic tasks, and also recent theoretical progress.
Chris Watkins obtained his BA and PhD at Cambridge in 1980 and 1989. His PhD thesis was on reinforcement learning. He has worked at Philips Research Labs, UK, for three fund management companies, and for a year at AT&T Bell Labs in New Jersey. He is currently a lecturer in computer science at Royal Holloway, University of London, and his research interests are in machine learning.
The Evening Lectures are free to both members and non-members of SGES.
For further information contact:
Dr. Hui Liu, Department of Computer Science, Birkbeck College hui@dcs.bbk.ac.uk