ResearchBenchmarksNoisy Action Robust MDP on Halfcheetah MuJoCo (high randomness)Follow1,765.71Worst-case ReturnARDT1,482.12281,555.74641,629.371,702.9936Jul 25, 2024Evaluation ResultsMethodMethodLinksWorst-case ReturnARDTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.071,765.71ESPERnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.071,699.8DTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.071,493.03