English

Finite-Sample Analysis of Off-Policy Natural Actor-Critic Algorithm

Machine Learning 2021-06-14 v2 Optimization and Control Machine Learning

Abstract

In this paper, we provide finite-sample convergence guarantees for an off-policy variant of the natural actor-critic (NAC) algorithm based on Importance Sampling. In particular, we show that the algorithm converges to a global optimal policy with a sample complexity of O(ϵ3log2(1/ϵ))\mathcal{O}(\epsilon^{-3}\log^2(1/\epsilon)) under an appropriate choice of stepsizes. In order to overcome the issue of large variance due to Importance Sampling, we propose the QQ-trace algorithm for the critic, which is inspired by the V-trace algorithm \cite{espeholt2018impala}. This enables us to explicitly control the bias and variance, and characterize the trade-off between them. As an advantage of off-policy sampling, a major feature of our result is that we do not need any additional assumptions, beyond the ergodicity of the Markov chain induced by the behavior policy.

Keywords

Cite

@article{arxiv.2102.09318,
  title  = {Finite-Sample Analysis of Off-Policy Natural Actor-Critic Algorithm},
  author = {Sajad Khodadadian and Zaiwei Chen and Siva Theja Maguluri},
  journal= {arXiv preprint arXiv:2102.09318},
  year   = {2021}
}
R2 v1 2026-06-23T23:17:09.526Z