Wasserstein Robust Reinforcement Learning

Reinforcement learning algorithms, though successful, tend to over-fit to training environments hampering their application to the real-world. This paper proposes WR$^{2}$L; a robust reinforcement learning algorithm with significant robust performance on low and high-dimensional control tasks. Our method formalises robust reinforcement learning as a novel min-max game with a Wasserstein constraint for a correct and convergent solver. Apart from the formulation, we also propose an efficient and scalable solver following a novel zero-order optimisation method that we believe can be useful to numerical optimisation in general. We contribute both theoretically and empirically. On the theory side, we prove that WR$^{2}$L converges to a stationary point in the general setting of continuous state and action spaces. Empirically, we demonstrate significant gains compared to standard and robust state-of-the-art algorithms on high-dimensional MuJuCo environments.