Wasserstein Flow Meets Replicator Dynamics: A Mean-Field Analysis of Representation Learning in Actor-Critic

arXiv:2112.13530v2 Announce Type: replace
Abstract: Actor-critic (AC) algorithms, empowered by neural networks, have had significant empirical success in recent years. However, most of the existing theoretical support for AC algorithms focuses on the case of linear function approximations, or linearized neural networks, where the feature representation is fixed throughout training. Such a limitation fails to capture the key aspect of representation learning in neural AC, which is pivotal in practical problems. In this work, we take a mean-field perspective on …

abstract actor actor-critic algorithms analysis arxiv case cs.lg dynamics feature flow function however linear math.oc mean networks neural networks representation representation learning stat.ml success support type