Worst-case time-constrained reinforcement learning on Humanoid MuJoCo (test)
4.02Normalized Worst-Case RewardTC-M2TD3
Evaluation results
| Method | Links | |
|---|---|---|
| TC-M2TD3Framework=TC2024.06 | 4.02 | |
| Oracle M2TD3Oracle=true2024.06 | 2.07 | |
| Oracle-TC-M2TD3Oracle=true, Framework=TC2024.06 | 1.69 | |
| Stacked-TC-RARLArchitecture=Stacked, Framework=TC2024.06 | 1.39 | |
| TC-RARLFramework=TC2024.06 | 1.25 | |
| M2TD3Framework=Robust2024.06 | 1 | |
| DRFramework=Domain Randomization2024.06 | 0.87 | |
| RARLFramework=Robust2024.06 | 0.8 | |
| Oracle-TC-RARLOracle=true, Framework=TC2024.06 | 0.19 | |
| TD3Framework=Vanilla2024.06 | 0 | |
| Stacked-TC-M2TD3Architecture=Stacked, Framework=TC2024.06 | -1.42 | |
| Oracle RARLOracle=true2024.06 | -2.8 |