Provably Efficient Representation Selection in Low-rank Markov Decision Processes: From Online to Offline RL

arXiv:2106.11935v2 Announce Type: replace
Abstract: The success of deep reinforcement learning (DRL) lies in its ability to learn a representation that is well-suited for the exploration and exploitation task. To understand how the choice of representation can improve the efficiency of reinforcement learning (RL), we study representation selection for a class of low-rank Markov Decision Processes (MDPs) where the transition kernel can be represented in a bilinear form. We propose an efficient algorithm, called ReLEX, for representation learning in both …

abstract arxiv cs.lg decision efficiency exploitation exploration learn lies low markov math.oc offline processes reinforcement reinforcement learning representation stat.ml study success type