Offline Reinforcement Learning on 70% Suboptimal Backward
49.6Final ReturnIQL + EG
Evaluation results
| Method | Links | |
|---|---|---|
| IQL + EGAlgorithm=IQL + EG, Number of seeds=52026.02 | 49.6 | |
| BC + EGAlgorithm=BC + EG, Number of seeds=52026.02 | 46.3 | |
| IQL + SELAlgorithm=IQL + SEL, Number of seeds=52026.02 | 44.45 | |
| BCAlgorithm=BC, Number of seeds=52026.02 | 41.42 | |
| IQL + PCGradAlgorithm=IQL + PCGrad, Number of seeds=52026.02 | 41.04 | |
| TD3+BC + EGAlgorithm=TD3+BC + EG, Number of seeds=52026.02 | 40.88 | |
| IQLAlgorithm=IQL, Number of seeds=52026.02 | 38.69 | |
| TD3+BCAlgorithm=TD3+BC, Number of seeds=52026.02 | 34.79 |