window-close on Meta-World v2 (test)
4,505Best Attack RewardPPO
Evaluation results
| Method | Links | |
|---|---|---|
| PPODefense mechanism=No defense2024.06 | 4,505 | |
| ATLA-PPODefense mechanism=AdvTraining2024.06 | 4,270 | |
| RAD-PPODefense mechanism=Regret2024.06 | 4,261 | |
| PA-ATLA-PPODefense mechanism=AdvTraining2024.06 | 4,041 | |
| WocaR-PPODefense mechanism=Partial smoothing2024.06 | 575 | |
| SA-PPODefense mechanism=Smoothing, Time-discounting=false2024.06 | 485 | |
| TDRT-PPODefense mechanism=Smoothing, Time-discounting=true2024.06 | 482 |