Sparse Feature Selection Makes Batch Reinforcement Learning More Sample
                 Efficient

          Botao Hao 1 Yaqi Duan 2 Tor Lattimore 1 Csaba Szepesvari 1 3 Mengdi Wang 2 1

             Abstract                  1. Introduction
                                   We consider batch reinforcement learning (RL), where the
   This paper provides a statistical analysis of high-       problem is to evaluate a target policy or to learn a good
   dimensional batch reinforcement learning ( ...