RAAM: The Benefits of Robustness in Approximating Aggregated MDPs in Reinforcement Learning

Open in new window