Optimistic Regret Bounds for Online Learning in Adversarial Markov Decision Processes

Open in new window