robotics
The Robot Model Had Learned the Operator’s Habit as Physics
Separating habit, shared dynamics and camera nuisance improved low-shot transfer across three robot datasets.
Summary
Separating habit, shared dynamics and camera nuisance improved low-shot transfer across three robot datasets.
Teleoperated demonstrations can look multimodal even when the executed action nearly determines the next physical state. The paper models three distinct causes: operator habit in choosing actions, shared physics after the action, and observation nuisance such as camera appearance. Its adaptation rule freezes a shared physics readout and updates only a thin interface. Across StackCube, DROID and RH20T, the split improved low-shot transfer and resisted corrupted adaptation data better than training from scratch, including multi-view pixel tests. The authors do not claim that every latent action is a human habit.
Why it matters
Separating habit, shared dynamics and camera nuisance improved low-shot transfer across three robot datasets.
Limits and context
- Its adaptation rule freezes a shared physics readout and updates only a thin interface.
- The authors do not claim that every latent action is a human habit.
Key claims
Separating habit, shared dynamics and camera nuisance improved low-shot transfer across three robot datasets.
Qualification: Its adaptation rule freezes a shared physics readout and updates only a thin interface.
Evidence: source-2026-09-10-008
Sources
- arXiv preprint 2609.09210arXiv · primary research
Corrections
No corrections have been recorded for this story.