Meta-Reinforcement Learning on MetaWorld ML-10
48.7ReachVariBAD
Evaluation Results
| Method | Links | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| VariBAD2025.12 | 48.7 | 15 | 7.8 | 50.1 | 100 | 79.5 | 10.3 | 79.7 | 37.7 | 17 | 26.1 | 2 | 0 | 21.3 | 0 | 45 | 10 | 33 | |
| CRAFT + State Decoderdecoder=State Decoder2025.12 | 41.3 | 49.1 | 22.7 | 81.2 | 99.5 | 94.6 | 17.5 | 99 | 56.1 | 23.8 | 52.9 | 1.4 | 0 | 36.4 | 0.4 | 58 | 18 | 45 | |
| CRAFT2025.12 | 40.1 | 55.5 | 39.4 | 79.8 | 100 | 91.8 | 53 | 82.1 | 77.4 | 66.4 | 42.8 | 20 | 0.7 | 30.3 | 0.7 | 69 | 19 | 52 | |
| SDVT2025.12 | 36.3 | 49.7 | 41 | 99.7 | 100 | 94.3 | 21 | 100 | 71.3 | 50.3 | 42.7 | 22.3 | 1.3 | 52.3 | 0 | 66 | 23.7 | 53 |