ResearchBenchmarksNoisy Action Robust MDP on Walker2D medium randomness MuJoCo Noisy Action Robust MDPFollow508.4Worst-case ReturnARDT386.304418.002449.7481.398Jul 25, 2024Evaluation ResultsMethodMethodLinksWorst-case ReturnARDTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.07508.4ESPERnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.07426.9DTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.07391