ResearchBenchmarksNoisy Action Robust MDP on Hopper MuJoCo (low randomness)Follow477.9Worst-case ReturnARDT50.356161.353272.35383.347Jul 25, 2024Evaluation ResultsMethodMethodLinksWorst-case ReturnARDTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.07477.9DTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.07112.1ESPERnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.0766.8