| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Future Frame Prediction | BridgeData 2 -> 10 frames | SSIM0.865 | 9 | |
| Value function estimation | BridgeData dt_rd_pnp ES & EM V2 | VOC Score74.7 | 7 | |
| Value function estimation | BridgeData dt_ft_stack ES & EM V2 | VOC69.8 | 7 | |
| Value function estimation | BridgeData dt_tk_stack Embodiment Shift V2 | VOC3.5 | 7 | |
| Value function estimation | BridgeData Embodiment Shift dt_tk_pnp V2 | VOC0.856 | 7 | |
| Value function estimation | BridgeData ms_sweep Environment Shift V2 | VOC0.49 | 7 | |
| Value function estimation | BridgeData Environment Shift V2 (rd_fold) | VOC72.6 | 7 | |
| Value function estimation | BridgeData Environment Shift V2 (ft_fold) | VOC69.3 | 7 | |
| Value function estimation | BridgeData Environment Shift V2 (td_fold) | VOC70.9 | 7 | |
| Value function estimation | BridgeData lm_pnp Environment Shift V2 | VOC72.5 | 7 | |
| Value function estimation | BridgeData tk_pnp In-Distribution V2 | VOC0.029 | 7 | |
| Expert vs. Non-Expert Trajectory Discrimination | BridgeData 5 scripted datasets V2 (in-distribution) | BinVOC1 | 7 | |
| Robot Failure Detection | BridgeData Fail V2 | Execution Accuracy85 | 7 | |
| Robot Manipulation | BridgeData WidowX V2 (full evaluation suite) | Pick and Place Success Rate90 | 6 | |
| Mistake detection | BridgeData V2 (test) | AP59.5 | 5 | |
| Robotic task reconstruction | BridgeData V2 (test) | FID191 | 4 | |
| Action Prediction | BridgeData V2 (test) | Top-1 Accuracy6.62 | 2 | |
| Goal-conditioned manipulation planning (All Tasks) | BridgeData | Metric- | 0 | |
| Goal-conditioned manipulation planning (Flip Cip) | BridgeData | Metric- | 0 | |
| Goal-conditioned manipulation planning (Close Drawer) | BridgeData | Metric- | 0 | |
| Goal-conditioned manipulation planning (Put in shelf) | BridgeData | Metric- | 0 | |
| Goal-conditioned manipulation planning (Put on plate) | BridgeData | Metric- | 0 |