Conservative and Greedy Approaches to Classification-based Policy Iteration - Archive ouverte HAL Access content directly
Conference Papers Year : 2012

Conservative and Greedy Approaches to Classification-based Policy Iteration

(1) , (1)
1
Mohammad Ghavamzadeh
  • Function : Author
  • PersonId : 868946
Alessandro Lazaric

Abstract

The existing classification-based policy iteration (CBPI) algorithms can be divided into two categories: {\em direct policy iteration} (DPI) methods that directly assign the output of the classifier (the approximate greedy policy w.r.t.~the current policy) to the next policy, and {\em conservative policy iteration} (CPI) methods in which the new policy is a mixture distribution of the current policy and the output of the classifier. The conservative policy update gives CPI a desirable feature, namely the guarantee that the policies generated by this algorithm improve at each iteration. We provide a detailed algorithmic and theoretical comparison of these two classes of CBPI algorithms. Our results reveal that in order to achieve the same level of accuracy, CPI requires more iterations, and thus, more samples than the DPI algorithm. Furthermore, CPI may converge to suboptimal policies whose performance is not better than DPI's.
Fichier principal
Vignette du fichier
Ghavamzadeh.pdf (133.7 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-00772610 , version 1 (10-01-2013)

Identifiers

  • HAL Id : hal-00772610 , version 1

Cite

Mohammad Ghavamzadeh, Alessandro Lazaric. Conservative and Greedy Approaches to Classification-based Policy Iteration. AAAI - 26th Conference on Artificial Intelligence, Jul 2012, Toronto, Canada. ⟨hal-00772610⟩
200 View
100 Download

Share

Gmail Facebook Twitter LinkedIn More