Continuous Control on MuJoCo v5
4,828Ant ScoreREDQ
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| REDQEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300 0002025.07 | 4,828 | 8,678 | 3,415 | 5,404 | 4,749 | |
| AEAEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300 0002025.07 | 4,744 | 9,661 | 3,555 | 5,376 | 4,652 | |
| AEAEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300,000, Number of seeds=102025.07 | 4,485 | 9,721 | 3,326 | 5,365 | 4,579 | |
| REDQEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300,000, Number of seeds=102025.07 | 4,473 | 8,692 | 3,169 | 4,718 | 4,697 | |
| AEAEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300 000, Number of seeds=102025.07 | 2,343 | 7,214 | 2,723 | 2,954 | 2,992 | |
| REDQEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300 000, Number of seeds=102025.07 | 2,303 | 6,722 | 2,843 | 3,031 | 3,389 | |
| DRNDEnsemble size=2, Update-to-Data Ratio=20, Interaction budget=300,000, Number of seeds=102025.07 | 819 | 7,893 | 3,097 | 2,726 | 5,053 | |
| DRNDEnsemble size=2, Update-to-Data Ratio=20, Interaction budget=300 0002025.07 | 797 | 7,892 | 3,015 | 2,853 | 5,054 | |
| DSACEnsemble size=2, Update-to-Data Ratio=20, Interaction budget=300 0002025.07 | 776 | 9,059 | 2,640 | 908 | 5,362 | |
| DSACEnsemble size=2, Update-to-Data Ratio=20, Interaction budget=300,000, Number of seeds=102025.07 | 731 | 9,059 | 2,680 | 1,051 | 5,363 | |
| DSACEnsemble size=2, Update-to-Data Ratio=20, Interaction budget=300 000, Number of seeds=102025.07 | 407 | 6,869 | 2,334 | 750 | 3,010 | |
| DRNDEnsemble size=2, Update-to-Data Ratio=20, Interaction budget=300 000, Number of seeds=102025.07 | 324 | 5,705 | 2,544 | 872 | 2,073 | |
| SAC-NEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300 000, Number of seeds=102025.07 | 33 | 4,791 | 1,318 | 2,357 | 975 | |
| SAC-NEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300 0002025.07 | 1 | 5,425 | 2,063 | 5,341 | 1,035 | |
| SAC-NEnsemble size=10, Update-to-Data Ratio=20, Interaction budget=300,000, Number of seeds=102025.07 | 1 | 5,586 | 1,992 | 4,903 | 1,581 |