Door Unlock on Meta-World v2 (test)
3,421Best Attack RewardPPO
Evaluation Results
| Method | Links | |
|---|---|---|
| PPODefense mechanism=No defense2024.06 | 3,421 | |
| ATLA-PPODefense mechanism=AdvTraining2024.06 | 3,277 | |
| PA-ATLA-PPODefense mechanism=AdvTraining2024.06 | 2,806 | |
| RAD-PPODefense mechanism=Regret2024.06 | 2,743 | |
| WocaR-PPODefense mechanism=Partial smoothing2024.06 | 1,073 | |
| SA-PPODefense mechanism=Smoothing, Time-discounting=false2024.06 | 787 | |
| TDRT-PPODefense mechanism=Smoothing, Time-discounting=true2024.06 | 691 |