Offline Reinforcement Learning on D4RL Hopper-medium-expert v2
116.5Normalized ReturnPhyB
Evaluation Results
| Method | Links | |
|---|---|---|
| PhyBModel category=Model-based2026.05 | 116.5 | |
| ADMModel category=Model-based2026.05 | 112.7 | |
| TD3+BCModel category=Model-free2026.05 | 112.2 | |
| PMDB2022.10 | 111.8 | |
| PMDBN=10, k=22022.10 | 111.8 | |
| PMDBModel category=Model-based2026.05 | 111.8 | |
| Uncertainty-Aware Rank-One MIMO Q Network2026.02 | 111.6 | |
| PBRL2026.02 | 110.8 | |
| EDAC2022.10 | 110.7 | |
| EDAC2022.10 | 110.7 | |
| BCQ2026.02 | 110.5 | |
| DMGModel category=Model-free2026.05 | 110.4 | |
| EPQModel category=Model-free2026.05 | 108.8 | |
| DT2022.02 | 107.6 | |
| DT2023.01 | 107.6 | |
| MOREL2022.10 | 105.6 | |
| MOReL2022.10 | 105.6 | |
| CQL2026.02 | 105.6 | |
| MOReLModel category=Model-based2026.05 | 105.6 | |
| CQL2022.02 | 105.4 | |
| CQL2023.01 | 105.4 | |
| Onestep2022.02 | 103.3 | |
| Onestep RL2023.01 | 103.3 | |
| SPOT2022.02 | 99.3 | |
| SPOT2023.01 | 99.3 | |
| TD3+BC2022.02 | 98 | |
| TD3+BC2023.01 | 98 | |
| TD3-BC2026.02 | 98 | |
| CQL2022.10 | 96.9 | |
| CQL2022.10 | 96.9 | |
| CQLModel category=Model-free2026.05 | 96.9 | |
| CPED2023.01 | 95.3 | |
| IQL2022.02 | 91.5 | |
| IQL2023.01 | 91.5 | |
| IQL2026.02 | 91.5 | |
| FQL+BCTraining timesteps=1M, Random seeds=102025.09 | 90.29 | |
| RAMBO2022.10 | 88.2 | |
| RAMBOModel category=Model-based2026.05 | 88.2 | |
| BRAC2022.10 | 81.3 | |
| BRAC2022.10 | 81.3 | |
| IQLTraining timesteps=1M, Random seeds=102025.09 | 73.16 | |
| BEAR2022.10 | 66.2 | |
| BEAR2022.10 | 66.2 | |
| MOPO2026.02 | 60.6 | |
| OTDFShift Type=Kinematic2025.12 | 58.1 | |
| EDAC-102026.02 | 58.1 | |
| AWAC2022.02 | 55.8 | |
| AWAC2023.01 | 55.8 | |
| DROCOShift Type=Kinematic2025.12 | 54 | |
| BC2022.10 | 53.9 | |
| BC2022.10 | 53.9 | |
| BC2022.02 | 52.5 | |
| BC2023.01 | 52.5 | |
| BEAR2026.02 | 50.6 | |
| BOSAShift Type=Kinematic2025.12 | 49.6 | |
| UWAV2026.02 | 44.9 | |
| FQLModel category=Model-free2026.05 | 43.8 | |
| DARAShift Type=Kinematic2025.12 | 17 | |
| IQLShift Type=Kinematic2025.12 | 12.5 | |
| IGDFShift Type=Kinematic2025.12 | 11.6 | |
| CQLShift Type=Kinematic2025.12 | 7.8 |