Problem Dependent Reinforcement Learning Bounds Which Can Identify Bandit Structure in MDPs.

Andrea Zanette, Emma Brunskill

Browse the full ICML paper archive.