Offline Reinforcement Learning on D4RL Adroit (expert, human)
46.1Adroit Door Return (Human)KFC++
Evaluation Results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| KFC++2021.11 | 46.1 | — | — | — | — | 0.2 | 60 | 9.4 | — | 68.4 | 4.2 | 5.6 | -0.1 | — | |
| KFC2021.11 | 44.1 | — | — | — | — | 0.2 | 61.3 | 7 | — | 71.3 | 3 | 3.6 | -0.1 | — | |
| S4RL-(Adv)Augmentation=Adversarial2021.11 | 35.3 | — | — | — | — | 0.2 | 51.2 | 6.3 | — | 58.2 | 2.9 | 0.8 | -0.1 | — | |
| S4RL-(N)Augmentation=Normal2021.11 | 27 | — | — | — | — | 0.2 | 44.4 | 5.9 | — | 57.1 | 2.7 | 2.1 | -0.1 | — | |
| EPQ2024.05 | 13.2 | — | — | — | — | 0.3 | 83.9 | 3.9 | — | 91.8 | 22.8 | 5.8 | 0.1 | 221.8 | |
| CABI+IQL2022.06 | 11.5 | 105.7 | 107.4 | 129.1 | 127.9 | 0.4 | 72 | 4.8 | 604.1 | 42.2 | 2.4 | 0.8 | -0.1 | — | |
| CQL2021.11 | 9.9 | — | — | — | — | 0.2 | 37.5 | 4.4 | — | 39.2 | 2.1 | 0.4 | -0.1 | — | |
| CQL2022.06 | 9.9 | 101.5 | 95 | 107 | 86.7 | 0.2 | 37.5 | 4.4 | 483.8 | 39.2 | 2.1 | 0.4 | -0.1 | — | |
| CQLsource=reproduced2024.05 | 9.9 | — | — | — | — | 0.2 | 37.5 | 4.4 | — | 39.2 | 2.1 | 0.4 | -0.1 | 93.6 | |
| CQLrandom_seeds=32021.06 | 9.1 | — | — | — | — | 0.35 | 55.8 | 2.1 | — | — | — | — | — | — | |
| CQLsource=original paper2024.05 | 9.1 | — | — | — | — | 0.4 | 55.8 | 2.1 | — | 40.3 | 5.7 | 3.5 | -0.1 | 116.8 | |
| ICQrandom_seeds=32021.06 | 6.4 | 103.9 | 109.5 | 123.8 | 128.3 | 1.5 | 91.3 | 2 | — | — | — | — | — | — | |
| Onestep2024.05 | 5.4 | — | — | — | — | 1.9 | 71.8 | 1.2 | — | 60 | 2.1 | 0.4 | -0.1 | 142.7 | |
| MISA2024.05 | 5.2 | — | — | — | — | 0.1 | 88.1 | 8.1 | — | 58.6 | 2.2 | 0.5 | -0.1 | 162.7 | |
| IQL2024.05 | 4.3 | — | — | — | — | 0.1 | 71.5 | 1.4 | — | 37.3 | 2.1 | 1.6 | -0.2 | 118.1 | |
| IQL2022.06 | 3.3 | 105.2 | 105.6 | 118.4 | 121.7 | 0 | 68.7 | 2.3 | 562.5 | 35.2 | 1.6 | 0.7 | -0.2 | — | |
| MCQ2024.05 | 2.3 | — | — | — | — | 0.1 | 68.5 | 0.3 | — | 49.4 | 1.4 | 1.3 | 0 | 123.3 | |
| UWAC2022.06 | 2.1 | 104.1 | 105.6 | 111.9 | 110.6 | 0.5 | 21.7 | 1.1 | 490.8 | 33.1 | 0.4 | 0 | -0.3 | — | |
| BC2024.05 | 2 | — | — | — | — | 0.1 | 63.9 | 1.2 | — | 37 | 0.6 | 0 | -0.3 | 104.5 | |
| BCrandom_seeds=32021.06 | 0.5 | 101.2 | 101.3 | 85.1 | 125.6 | 0 | 34.4 | 1.5 | — | — | — | — | — | — | |
| BC2022.06 | 0.5 | 34.9 | 101.3 | 85.1 | 125.6 | 0 | 34.4 | 1.5 | 440.8 | 56.9 | 0.8 | -0.1 | -0.1 | — | |
| AWRrandom_seeds=32021.06 | 0.4 | 102.9 | 91.5 | 111 | 39 | 0 | 12.3 | 1.2 | — | — | — | — | — | — | |
| AWR2022.06 | 0.4 | 102.9 | 91.5 | 111 | 39 | 0 | 12.3 | 1.2 | 386.5 | 28 | 0.4 | 0 | -0.2 | — | |
| BCQrandom_seeds=32021.06 | 0 | 99 | 41.6 | 114.9 | 107.2 | -0.1 | 68.9 | 0.5 | — | — | — | — | — | — | |
| TD3+BC2024.05 | 0 | — | — | — | — | 0.1 | 64.8 | 1.8 | — | 49 | 0.2 | 0 | -0.2 | 115.7 | |
| MOPO2022.06 | -0.2 | -0.2 | -0.2 | -0.6 | 0.3 | -0.3 | 9.7 | 0.2 | 6.4 | -2.1 | 0.2 | -0.1 | -0.3 | — | |
| BRAC-prandom_seeds=32021.06 | -0.3 | -0.3 | -0.3 | -3.5 | 0.3 | -0.3 | 8.1 | 0.3 | — | — | — | — | — | — | |
| BEAR2022.06 | -0.3 | 103.4 | 98.6 | 105.9 | 127.3 | -0.3 | -1 | 0.3 | 460.3 | 26.5 | 0.3 | -0.1 | -0.3 | — | |
| COMBO2022.06 | -0.3 | 4.9 | 17.2 | 11.5 | 0.3 | -0.3 | 27.7 | 0.2 | 59.1 | -2.4 | 0.4 | 0 | -0.1 | — |