Offline Reinforcement Learning on halfcheetah medium
68.2Normalized ScoreAnti-exploration Method
Evaluation Results
| Method | Links | |
|---|---|---|
| Anti-exploration Method2026.02 | 68.2 | |
| SAC-RND2026.02 | 66.6 | |
| Q-ALIGN DT2026.05 | 65.3 | |
| GPC-SAC2026.02 | 60.8 | |
| QCS2026.05 | 59 | |
| ProposedSeeds=52023.10 | 58.1 | |
| DMG2026.02 | 54.9 | |
| COMBOSeeds=52023.10 | 54.2 | |
| ATACSeeds=52023.10 | 53.3 | |
| QT2026.05 | 51.4 | |
| TD3+BC2026.05 | 48.3 | |
| IQLSeeds=52023.10 | 47.4 | |
| O-DICE2026.02 | 47.4 | |
| IQL2026.02 | 47.4 | |
| IQL2026.05 | 47.4 | |
| CQL2026.02 | 47 | |
| CQLReward Specification=Step-wise2024.02 | 44.4 | |
| LSDT2026.05 | 43.6 | |
| DM2026.05 | 43.5 | |
| TD3-CVAE2026.02 | 43.2 | |
| CGDT2026.05 | 43 | |
| DC2026.05 | 43 | |
| DT2026.05 | 42.6 | |
| DTReward Specification=Final Return2024.02 | 42.4 | |
| QDTReward Specification=Final Return2024.02 | 42.4 | |
| STCDynamics Shift=Gravity, Target Quality=medium2026.02 | 42.4 | |
| QDTReward Specification=Step-wise2024.02 | 42.3 | |
| DTReward Specification=Step-wise2024.02 | 42.1 | |
| BEARSeeds=52023.10 | 41.7 | |
| DARADynamics Shift=Gravity, Target Quality=medium2026.02 | 41.2 | |
| BCQSeeds=52023.10 | 40.7 | |
| OTDFDynamics Shift=Gravity, Target Quality=medium2026.02 | 40.7 | |
| IQLDynamics Shift=Gravity, Target Quality=medium2026.02 | 39.6 | |
| BOSADynamics Shift=Gravity, Target Quality=medium2026.02 | 38.9 | |
| OptiDICESeeds=52023.10 | 38.2 | |
| CQLSeeds=52023.10 | 37.2 | |
| SRPODynamics Shift=Gravity, Target Quality=medium2026.02 | 36.9 | |
| IGDFDynamics Shift=Gravity, Target Quality=medium2026.02 | 36.6 | |
| TD3+BCSeeds=52023.10 | 27.8 | |
| DVDF-IGDF2025.12 | 26.7 | |
| OTDF2025.12 | 24.6 | |
| DVDF-OTDF2025.12 | 22.9 | |
| IGDF2025.12 | 22.6 | |
| BCQ2026.02 | 20.23 | |
| AWAC2026.02 | 19.69 | |
| IQL2025.12 | 18.7 | |
| BOSA2025.12 | 17.3 | |
| DARA2025.12 | 16.1 | |
| QDFM2026.02 | 4.92 | |
| AWBC2026.02 | 2.22 | |
| CQLReward Specification=Final Return2024.02 | 1 | |
| GreedyQ2026.02 | -0.23 | |
| CQL2026.02 | -0.58 |