Matthew E. Taylor's Publications

Sorted by DateClassified by Publication TypeSorted by First Author Last NameClassified by Research Category

Generalized Domains for Empirical Evaluations in Reinforcement Learning

Shimon Whiteson, Brian Tanner, Matthew E. Taylor, and Peter Stone. Generalized Domains for Empirical Evaluations in Reinforcement Learning. In Proceedings of the Fourth Workshop on Evaluation Methods for Machine Learning at ICML-09, June 2009.
Fourth annual workshop on Evaluation Methods for Machine Learning

Download

[PDF]90.2kB  [postscript]209.0kB  

Abstract

Many empirical results in reinforcement learning are based on a very small set of environments. These results often represent the best algorithm parameters that were found after an ad-hoc tuning or fitting process. We argue that presenting tuned scores from a small set of environments leads to method overfitting, wherein results may not generalize to similar environments. To address this problem, we advocate empirical evaluations using generalized domains: parameterized problem generators that explicitly encode variations in the environment to which the learner should be robust. We argue that evaluating across a set of these generated problems offers a more meaningful evaluation of reinforcement learning algorithms.

BibTeX Entry

@inproceedings(ICMLWS09-Whiteson,
  author="Shimon Whiteson and Brian Tanner and Matthew E.\ Taylor and Peter Stone",
  title="Generalized Domains for Empirical Evaluations in Reinforcement Learning",
  Booktitle="Proceedings of the Fourth Workshop on Evaluation Methods for Machine Learning at {ICML}-09",
  month="June",
  year= "2009",
  wwwnote={<a
  href="http://www.site.uottawa.ca/ICML09WS/">Fourth annual workshop on Evaluation Methods for Machine Learning</a>},
  abstract = {Many empirical results in reinforcement learning are
  based on a very small set of environments.  These results often
  represent the best algorithm parameters that were found after an
  ad-hoc tuning or fitting process.  We argue that presenting tuned
  scores from a small set of environments leads to method overfitting,
  wherein results may not generalize to similar environments.  To
  address this problem, we advocate empirical evaluations using
  generalized domains: parameterized problem generators that
  explicitly encode variations in the environment to which the learner
  should be robust.  We argue that evaluating across a set of these
  generated problems offers a more meaningful evaluation of
  reinforcement learning algorithms.},
)

Generated by bib2html.pl (written by Patrick Riley ) on Mon Apr 19, 2010 14:12:46