Robotic Manipulation on adroit-pen (OSR, SR)
98.7Success Rate (SR)Cal-QL
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Cal-QL2026.05 | 98.7 | 75 | — | — | — | — | |
| RankQ2026.05 | 96.9 | 5 | — | — | — | — | |
| RankQ+SAC2026.05 | 91.2 | 5 | — | — | — | — | |
| Cal-QL+SAC2026.05 | 83.3 | 75 | — | — | — | — | |
| CQL2026.05 | 81.6 | 53.3 | — | — | — | — | |
| Hybrid RL2026.05 | 73.8 | 3.3 | — | — | — | — | |
| SAC+OFF2026.05 | 73.7 | 3.3 | — | — | — | — | |
| CQL+SAC2026.05 | 71.1 | 53.3 | — | — | — | — | |
| SAC2026.05 | 48.8 | 13.3 | — | — | — | — | |
| HiMAQ+SACBaseline Algorithm=SAC, Macro Action Length (H)=82026.05 | 0.51 | — | 0.58 | 0.58 | 0.61 | 0.67 | |
| MAQ+SACBaseline Algorithm=SAC, Macro Action Length (H)=82026.05 | 0.41 | — | 0.58 | 0.58 | 0.61 | 0.67 | |
| BCBaseline Algorithm=Behavioral Cloning, Macro Action Length (H)=12026.05 | 0.4 | — | 0.53 | 0.58 | 0.59 | 0.65 | |
| SACBaseline Algorithm=SAC, Macro Action Length (H)=12026.05 | 0.32 | — | 0.06 | -0.34 | 0.29 | 0.22 |