Abstract
Vision-language-action policies benefit from geometric supervision, but current-frame geometry alone does not explicitly describe the changes associated with manipulation. This design is motivated by the goal of learning an embodiment-agnostic visual interface that can be pretrained across robot and egocentric video before robot-specific action alignment. We introduce Geometry-Change VLA (GC-VLA), which learns to predict multiview future-current geometry-change tokens from current observations. Offline frame pairs define a nominal 0.5-second prediction horizon; future observations are used only to construct training targets. Stage 1 trains a geometry-change vision-language model (GC-VLM). Stage 2 introduces a continuous ActionExpert and aligns it with robot actions while stopping action-flow gradients at the VLM interface. Stage 3 enables these gradients to update the trainable VLM components jointly with the ActionExpert. Stage 4 freezes GC-VLA and applies Geometry-Conditioned Residual Flow (GCRF), using a binary intervention router and a single bounded residual velocity policy learned from closed-loop feedback. GC-VLA achieves 95.20% success on LIBERO, and GC-VLA with GCRF achieves 99.55%. Inference uses current observations and the learned GC representation without executing the offline target encoders.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Pahk, J., Kang, J., Park, T., An, J., Kimm, S. M., Kim, J. J., & Zhang, B. T. (2026). HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery. https://omanscience.com/en/articles/habilis-brain-0-geometry-change-supervision-for-vision-language-action-and-residual-flow-recovery
MLA 9
Pahk, Jinu, et al. "HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery." https://omanscience.com/en/articles/habilis-brain-0-geometry-change-supervision-for-vision-language-action-and-residual-flow-recovery.
Chicago (author–date)
Pahk, Jinu, Jesoon Kang, Taegeon Park, Jisu An, Soo Min Kimm, Jae-Joon Kim, and Byoung-Tak Zhang. 2026. "HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery." https://omanscience.com/en/articles/habilis-brain-0-geometry-change-supervision-for-vision-language-action-and-residual-flow-recovery.
Harvard
Pahk, J., Kang, J., Park, T., An, J., Kimm, S. M., Kim, J. J. and Zhang, B. T. (2026) 'HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery', Available at: https://omanscience.com/en/articles/habilis-brain-0-geometry-change-supervision-for-vision-language-action-and-residual-flow-recovery.
Vancouver
Pahk J, Kang J, Park T, An J, Kimm SM, Kim JJ, et al. HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery. https://omanscience.com/en/articles/habilis-brain-0-geometry-change-supervision-for-vision-language-action-and-residual-flow-recovery
IEEE
J. Pahk, J. Kang, T. Park, J. An, S. M. Kimm, J. J. Kim, and B. T. Zhang, "HABILIS Brain 0: Geometry-Change Supervision for Vision-Language-Action and Residual Flow Recovery," https://omanscience.com/en/articles/habilis-brain-0-geometry-change-supervision-for-vision-language-action-and-residual-flow-recovery.