Offline Reinforcement Learning (Navigation) on AntMaze umaze-diverse D4RL (ud)
94Expert Normalized ReturnDMM
Evaluation Results
| Method | Links | |
|---|---|---|
| DMMArchitecture=SSM-based2026.02 | 94 | |
| CQLArchitecture=Value-based2026.02 | 84 | |
| DCArchitecture=Transformer-based2026.02 | 78.5 | |
| WTArchitecture=Transformer-based2026.02 | 71.5 | |
| TD3+BCArchitecture=Value-based2026.02 | 71.4 | |
| DS4Architecture=SSM-based2026.02 | 64.6 | |
| IQLArchitecture=Value-based2026.02 | 62.2 | |
| QLDTArchitecture=Transformer-based2026.02 | 62.2 | |
| DMArchitecture=SSM-based2026.02 | 62 | |
| DTArchitecture=Transformer-based2026.02 | 42.2 |