Paper published in a book (Scientific congresses and symposiums)
Selecting concise sets of samples for a reinforcement learning agent
Ernst, Damien
2005In Proceedings of the 3rd International Conference on Computational Intelligence, Robotics and Autonomous Systems (CIRAS 2005)
Peer reviewed
 

Files


Full Text
ernst-ciras2005.pdf
Publisher postprint (1.05 MB)
Download
Annexes
concise-tech-00.pdf
(1.08 MB)
Download
ernst-ciras2005-slides.pdf
(361.97 kB)
Download

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
reinforcement learning; fitted Q iteration; concise sets
Abstract :
[en] We derive an algorithm for selecting from the set of samples gathered by a reinforcement learning agent interacting with a deterministic environment, a concise set from which the agent can extract a good policy. The reinforcement learning agent is assumed to extract policies from sets of samples by solving a sequence of standard supervised learning regression problems. To identify concise sets, we adopt a criterion based on an error function defined from the sequence of models produced by the supervised learning algorithm. We evaluate our approach on two-dimensional maze problems and show its good performances when problems are continuous.
Disciplines :
Computer science
Author, co-author :
Ernst, Damien  ;  Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Systèmes et modélisation
Language :
English
Title :
Selecting concise sets of samples for a reinforcement learning agent
Publication date :
2005
Event name :
3rd International Conference on Computational Intelligence, Robotics and Autonomous Systems (CIRAS 2005)
Event place :
Singapore, Singapore
Event date :
22-26 August 2005
Audience :
International
Main work title :
Proceedings of the 3rd International Conference on Computational Intelligence, Robotics and Autonomous Systems (CIRAS 2005)
Peer reviewed :
Peer reviewed
Funders :
F.R.S.-FNRS - Fonds de la Recherche Scientifique [BE]
Available on ORBi :
since 27 May 2009

Statistics


Number of views
109 (17 by ULiège)
Number of downloads
100 (8 by ULiège)

Bibliography


Similar publications



Contact ORBi