Approximate Solutions to Markov Decision ProcessesRegret bounds for prediction problemsStable fitted reinforcement learning
