Offline Reinforcement Learning on D4RL hopper-medium-expert
119.2Normalized ScoreATAC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ATACSeeds=52023.10 | 119.2 | — | |
| ProposedSeeds=52023.10 | 117.8 | — | |
| NEUBAYMethod Category=Ours2025.12 | 114.8 | — | |
| VIPOMethod Category=Conservative model-based2025.12 | 113.2 | — | |
| PSPOAveraged across=4 random seeds2026.05 | 112.8 | — | |
| ADMPOMethod Category=Conservative model-based2025.12 | 112.7 | — | |
| AMGNumber of random seeds=42026.03 | 112.7 | — | |
| ADM2026.05 | 112.7 | — | |
| MOBILEMethod Category=Conservative model-based2025.12 | 112.6 | — | |
| RATE2023.06 | 112.5 | — | |
| TD3+BCSeeds=52023.10 | 112.4 | — | |
| CABI+TD3-BCMethod Type=Model-based2023.10 | 112.4 | — | |
| BC(CQL)Evaluation=Average score over 100 episodes2024.10 | 111.9 | — | |
| Fu et al.Algorithmic Template=Iterative2021.06 | 111.9 | — | |
| DMambasource=Lv et al., 20242023.06 | 111.9 | — | |
| RRPINumber of random seeds=42026.03 | 111.9 | — | |
| ATACMethod Type=Model-free2023.10 | 111.9 | — | |
| DDImplementation=Original2023.06 | 111.8 | — | |
| DDCategory=Diffusion Planner2024.05 | 111.8 | — | |
| Decision Diffuser2025.12 | 111.8 | — | |
| PMDBNumber of random seeds=42026.03 | 111.8 | — | |
| PMDBMethod Type=Model-based2023.10 | 111.8 | — | |
| PMDB2026.05 | 111.8 | — | |
| DDImplementation=reproduced2023.06 | 111.6 | — | |
| AdaptDiffuserCategory=Diffusion Planner2024.05 | 111.6 | — | |
| CBOPMethod Category=Bayesian-inspired2025.12 | 111.6 | — | |
| OptiDICESeeds=52023.10 | 111.5 | — | |
| CQLSeeds=52023.10 | 111.4 | — | |
| ROMI-CQL2021.10 | 111.4 | — | |
| Count-MORLMethod Type=Model-based2023.10 | 111.4 | — | |
| COMBOSeeds=52023.10 | 111.1 | — | |
| COMBO2021.10 | 111.1 | — | |
| COMBOCategory=Model-based RL2024.05 | 111.1 | — | |
| COMBOMethod Category=Model-free2025.12 | 111.1 | — | |
| COMBO2025.12 | 111.1 | — | |
| DMambasource=Ota, 20242023.06 | 111.1 | — | |
| DQLNFE=52026.03 | 111.1 | — | |
| COMBOMethod Type=Model-based2023.10 | 111.1 | — | |
| DMModel Category=SSM-based2026.02 | 111 | — | |
| BCQSeeds=52023.10 | 110.9 | — | |
| Decision Stacks2023.06 | 110.9 | — | |
| DS4Model Category=SSM-based2026.02 | 110.8 | — | |
| EDACMethod Category=Model-free2025.12 | 110.7 | — | |
| EDACMethod Type=Model-free2023.10 | 110.7 | — | |
| MambaDM2023.06 | 110.5 | — | |
| MPDiffuser+Rankranking=true2025.12 | 110.4 | — | |
| DMGNumber of random seeds=42026.03 | 110.4 | — | |
| DMG2026.05 | 110.4 | — | |
| DMMModel Category=SSM-based2026.02 | 110.2 | — | |
| DCModel Category=SSM-based2026.02 | 110.1 | — | |
| Trajectory Transformerdiscretization=quantile2021.06 | 110 | — | |
| TT2023.06 | 110 | — | |
| TT2023.06 | 110 | — | |
| TT2026.03 | 110 | — | |
| MPDiffuserranking=false2025.12 | 109.9 | — | |
| Planner2025.12 | 109.5 | — | |
| LEQMethod Category=Conservative model-based2025.12 | 109.4 | — | |
| MoMoMethod Category=Conservative model-based2025.12 | 109.1 | — | |
| CAMPAccess=Scripted preferences2024.04 | 108.9 | — | |
| EPQNumber of random seeds=42026.03 | 108.8 | — | |
| EPQ2026.05 | 108.8 | — | |
| MOREL2021.10 | 108.7 | — | |
| MOREL2023.06 | 108.7 | — | |
| MORELCategory=Model-based RL2024.05 | 108.7 | — | |
| DTModel Category=Transformer-based2026.02 | 108.7 | — | |
| MOReL2026.03 | 108.7 | — | |
| OPPOAccess=Scripted preferences, Implementation=Reported2024.04 | 108 | — | |
| SUMOMethod Category=Conservative model-based2025.12 | 107.8 | — | |
| Decision Transformer2021.06 | 107.6 | — | |
| DT2023.06 | 107.6 | — | |
| DT2025.12 | 107.6 | — | |
| DT2023.06 | 107.6 | — | |
| DT2026.03 | 107.6 | — | |
| CROPMethod Type=Model-based2023.10 | 107.5 | — | |
| IQLAccess=True reward functions2024.04 | 107.4 | — | |
| Diffuser2023.06 | 107.2 | — | |
| Diffuser2025.12 | 107.2 | — | |
| Decision Stacks2023.06 | 107 | — | |
| CQL2021.10 | 106.8 | — | |
| BAIL2021.10 | 106.4 | — | |
| Trajectory Transformerdiscretization=uniform2021.06 | 106 | — | |
| APE-VMethod Category=Conservative model-based2025.12 | 105.7 | — | |
| MOReLNumber of random seeds=42026.03 | 105.6 | — | |
| MOReL2026.05 | 105.6 | — | |
| TAP2023.06 | 105.5 | — | |
| CQL2021.06 | 105.4 | — | |
| CQL2023.06 | 105.4 | — | |
| CQLModel Category=Value-based2026.02 | 105.4 | — | |
| CQL2023.06 | 105.4 | — | |
| CQL2026.03 | 105.4 | — | |
| IDQL2025.12 | 105.3 | — | |
| DT2023.06 | 105.2 | — | |
| EDTModel Category=Transformer-based2026.02 | 104.7 | — | |
| KDPNFE=12026.03 | 103.6 | — | |
| DiffuserCategory=Diffusion Planner2024.05 | 103.3 | — | |
| DiffuserNFE=202026.03 | 103.3 | — | |
| MROracle status=false, Implementation source=reimpl.2023.05 | 103 | — | |
| Rev. KL Reg.Algorithmic Template=One-step2021.06 | 102.1 | — | |
| CACNFE=22026.03 | 100.4 | — | |
| OPPOAccess=Scripted preferences, Implementation=Re-implementation2024.04 | 99.5 | — |