Improved Regret Bound and Experience Replay in Regularized Policy Iteration.

Nevena Lazic, Dong Yin, Yasin Abbasi-Yadkori, Csaba Szepesvri

Browse the full ICML paper archive.