ResearchBenchmarksLocomotion on MuJoCo HalfCheetahFollow3,688,291Target RewardPPO3,503,876.453,596,083.7253,688,2913,780,498.275Jun 6, 2024Evaluation ResultsMethodMethodLinksTarget RewardClean RewardTargeted PGD RewardRew Max (SA-RL) RewardBIA-ILfD RewardAverage Attack RewardPPODefense=No defenseDefense=No defense2024.063,688,29121,7842,823,519402,398360,2753,482TDRT-PPODefense=TDRTDefense=TDRT2024.063,688,291197,0382,023,274260,648208,4392,237