ResearchBenchmarksNoisy Action Robust MDP on Halfcheetah MuJoCo (low randomness)Follow1,691.8Worst-case ReturnESPER1,404.9681,479.4341,553.91,628.366Jul 25, 2024Evaluation ResultsMethodMethodLinksWorst-case ReturnESPERnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.071,691.8ARDTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.071,509.2DTnumber of seeds=5, adv...number of seeds=5, adversarial policies=8, target returns=best2024.071,416