handle-pull-side on Meta-World v2 (test)
4,268Best Attack RewardPPO
Evaluation Results
| Method | Links | |
|---|---|---|
| PPODefense mechanism=No defense2024.06 | 4,268 | |
| RAD-PPODefense mechanism=Regret2024.06 | 1,086 | |
| ATLA-PPODefense mechanism=AdvTraining2024.06 | 532 | |
| PA-ATLA-PPODefense mechanism=AdvTraining2024.06 | 512 | |
| WocaR-PPODefense mechanism=Partial smoothing2024.06 | 33 | |
| SA-PPODefense mechanism=Smoothing, Time-discounting=false2024.06 | 10 | |
| TDRT-PPODefense mechanism=Smoothing, Time-discounting=true2024.06 | 7 |