Off-policy Actor-Critic algorithms have demonstrated phenomenal experimental performance but still require better explanations. To this end, we show its policy evaluation error on the distribution of transitions decomposes into: a Bellman error, a bias from policy mismatch, and a variance term from sampling. By comparing the magnitude of bias and variance, we explain the success of the Emphasizing Recent Experience sampling and 1/age weighted sampling. Both sampling strategies yield smaller bias and variance and are hence preferable to uniform sampling.
@article{arxiv.2110.02421,
title = {Explaining Off-Policy Actor-Critic From A Bias-Variance Perspective},
author = {Ting-Han Fan and Peter J. Ramadge},
journal= {arXiv preprint arXiv:2110.02421},
year = {2021}
}