Offline Reinforcement Learning on D4RL hopper-random
53.6Normalized ScoreMOREL
Evaluation Results
| Method | Links | |
|---|---|---|
| MOREL2021.10 | 53.6 | |
| MOREL2021.02 | 53.6 | |
| MOReLNumber of random seeds=42026.03 | 38.1 | |
| MOReL2026.05 | 38.1 | |
| RRPINumber of random seeds=42026.03 | 35 | |
| PMDBNumber of random seeds=42026.03 | 32.7 | |
| AMGNumber of random seeds=42026.03 | 32.7 | |
| PMDBMethod Type=Model-based2023.10 | 32.7 | |
| PMDB2026.05 | 32.7 | |
| ADM2026.05 | 32.7 | |
| EPQNumber of random seeds=42026.03 | 32.1 | |
| EPQ2026.05 | 32.1 | |
| PSPOAveraged across=4 random seeds2026.05 | 31.9 | |
| CROPMethod Type=Model-based2023.10 | 31.8 | |
| Repb-SDE2021.10 | 31.7 | |
| Repb-SDEalgorithm class=model-based2021.10 | 31.7 | |
| BEAR2021.10 | 31.4 | |
| CQLalgorithm class=model-free2021.10 | 31.4 | |
| SAC-NImplementation=Ours2021.10 | 31.3 | |
| Count-MORLMethod Type=Model-based2023.10 | 30.7 | |
| ROMI-CQL2021.10 | 30.2 | |
| ROMI-CQLbase algorithm=CQL2021.10 | 30.2 | |
| RAMBONumber of random seeds=42026.03 | 25.4 | |
| RAMBO2026.05 | 25.4 | |
| EDACImplementation=Ours2021.10 | 25.3 | |
| EDACMethod Type=Model-free2023.10 | 25.3 | |
| DMGNumber of random seeds=42026.03 | 20.4 | |
| DMG2026.05 | 20.4 | |
| ProposedSeeds=52023.10 | 18.7 | |
| COMBOSeeds=52023.10 | 17.9 | |
| COMBO2021.10 | 17.9 | |
| COMBO2021.02 | 17.9 | |
| ATACSeeds=52023.10 | 17.5 | |
| ATACMethod Type=Model-free2023.10 | 17.5 | |
| Fu et al.Algorithmic Template=Iterative2021.06 | 12.2 | |
| BRAC-v2021.02 | 12.2 | |
| CABI+TD3-BCMethod Type=Model-based2023.10 | 11.9 | |
| MOPO2021.02 | 11.7 | |
| RAMBOMethod Type=Model-based2023.10 | 11.5 | |
| BEARSeeds=52023.10 | 11.4 | |
| Fisher-BRCimplementation=author-provided (re-run)2021.06 | 11.4 | |
| BEAR2021.02 | 11.4 | |
| SAC-off2021.02 | 11.3 | |
| OptiDICESeeds=52023.10 | 11.2 | |
| TD3+BCSeeds=52023.10 | 11.1 | |
| BRAC-pimplementation=reported2021.06 | 11.1 | |
| TD3+BCalpha=2.52021.06 | 11 | |
| BRAC-p2021.02 | 11 | |
| IQLSeeds=52023.10 | 10.8 | |
| CQLImplementation=Original Paper2021.10 | 10.8 | |
| CQL2021.02 | 10.8 | |
| CQLSeeds=52023.10 | 10.7 | |
| CQLimplementation=author-provided (re-run)2021.06 | 10.7 | |
| BCQSeeds=52023.10 | 10.6 | |
| BRAC-P2021.10 | 10 | |
| SACImplementation=Reproduced2021.10 | 9.9 | |
| AWACimplementation=reported2021.06 | 9.6 | |
| BCimplementation=own2021.06 | 9.5 | |
| BRACImplementation=Reproduced2021.10 | 8.1 | |
| ROMI-BCQ2021.10 | 8 | |
| BAIL2021.10 | 8 | |
| BCQ2021.10 | 7.9 | |
| BCQImplementation=Reproduced2021.10 | 7.8 | |
| Rev. KL Reg.Algorithmic Template=One-step2021.06 | 7.8 | |
| BRAC-v2021.10 | 7.1 | |
| COMBOMethod Type=Model-based2023.10 | 7 | |
| BooT-oGeneration Strategy=Autoregressive2022.06 | 6.8 | |
| CQL2021.10 | 6.7 | |
| Easy BCQAlgorithmic Template=One-step2021.06 | 6.7 | |
| TT2022.06 | 6.7 | |
| BooT-rGeneration Strategy=Teacher-forcing2022.06 | 6.6 | |
| BooT-rGeneration Strategy=Autoregressive2022.06 | 6.5 | |
| BooT-oGeneration Strategy=Teacher-forcing2022.06 | 6.5 | |
| Exp. WeightAlgorithmic Template=One-step2021.06 | 5.6 | |
| CQLImplementation=Reproduced2021.10 | 5.3 | |
| CQLNumber of random seeds=42026.03 | 5.3 | |
| CQL2026.05 | 5.3 | |
| MOPO2021.10 | 4.1 | |
| BCImplementation=Reproduced2021.10 | 3.7 | |
| BEARImplementation=Reproduced2021.10 | 3.6 | |
| REMImplementation=Reproduced2021.10 | 3.4 | |
| UWACImplementation=Reproduced2021.10 | 2.6 | |
| BCAlgorithmic Template=One-step2021.06 | 2.6 | |
| BC2021.02 | 1.6 | |
| BC2021.10 | 0.9 | |
| BC2021.10 | 0.9 |