Improved Regret Bound and Experience Replay in Regularized Policy Iteration

Open in new window