Delay-Adapted Policy Optimization and Improved Regret for Adversarial MDP with Delayed Bandit Feedback.

Tal Lancewicki, Aviv Rosenberg, Dmitry Sotnikov

Browse the full ICML paper archive.