English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
  Efficient data reuse in value function approximation

Hachiya, H., Akiyama, T., Sugiyama, M., & Peters, J. (2009). Efficient data reuse in value function approximation. In 2009 IEEE Symposium on Adaptive Dynamic Programming and Reinforcement Learning (pp. 8-15). Piscataway, NJ, USA: IEEE Service Center.

Item is

Files

show Files

Locators

show
hide
Description:
-

Creators

show
hide
 Creators:
Hachiya, H, Author              
Akiyama, T, Author
Sugiyama, M, Author
Peters, J1, 2, Author              
Affiliations:
1Department Empirical Inference, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_1497795              
2Max Planck Institute for Biological Cybernetics, Max Planck Society, Spemannstrasse 38, 72076 Tübingen, DE, ou_1497794              

Content

show
hide
Free keywords: -
 Abstract: Off-policy reinforcement learning is aimed at efficiently using data samples gathered from a policy that is different from the currently optimized policy. A common approach is to use importance sampling techniques for compensating for the bias of value function estimators caused by the difference between the data-sampling policy and the target policy. However, existing off-policy methods often do not take the variance of the value function estimators explicitly into account and therefore their performance tends to be unstable. To cope with this problem, we propose using an adaptive importance sampling technique which allows us to actively control the trade-off between bias and variance. We further provide a method for optimally determining the trade-off parameter based on a variant of cross-validation. The usefulness of the proposed approach is demonstrated through simulated swing-up inverted-pendulum problem.

Details

show
hide
Language(s):
 Dates: 2009-05
 Publication Status: Published in print
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: DOI: 10.1109/ADPRL.2009.4927519
BibTex Citekey: 5771
 Degree: -

Event

show
hide
Title: 2009 IEEE International Symposium on Adaptive Dynamic Programming and Reinforcement Learning (ADPRL 2009)
Place of Event: Nashville, TN, USA
Start-/End Date: 2009-03-30 - 2009-04-02

Legal Case

show

Project information

show

Source 1

show
hide
Title: 2009 IEEE Symposium on Adaptive Dynamic Programming and Reinforcement Learning
Source Genre: Proceedings
 Creator(s):
Affiliations:
Publ. Info: Piscataway, NJ, USA : IEEE Service Center
Pages: - Volume / Issue: - Sequence Number: - Start / End Page: 8 - 15 Identifier: ISBN: 978-1-4244-2761-1