Geometry-Change VLA (GC-VLA), which learns to predict multiview future-current geometry-change tokens from current observations and applies Geometry-Conditioned Residual Flow (GCRF), using a binary intervention router and a single bounded residual velocity policy learned from closed-loop feedback.
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.
Vision-language-action (VLA) models encode semantic information from vision-language pretraining, but manipulation also requires precise spatial reasoning. We present GeoBridge-VLA, a two-stage method for learning geometric features from a pretrained VLA's frozen visual encoder and using them for action prediction. Sta...
Hyunjung Song, Kangmin Kim, Loren Jinsoo Um et al.· 0 citations
Dynin-Robotics achieves competitive performance on LIBERO and zero-shot LIBERO-Plus, together with a 78.4% average success rate across four manipulation conditions on a Franka Research 3 robot.
Hoe-hwan Lee, Jaeik Kim, Jusang Oh et al.· arXiv.org· 0 citations
This work proposes SAVLA, an end-to-end symmetry-aware VLA model for robust and data-efficient policy learning, which keeps the pretrained vision-language backbone entirely frozen while combining it with an equivariant flow-matching action head and a learned canonicalizer.
Vision-Language-Action (VLA) models have recently incorporated world models to provide richer dynamic supervision beyond sparse action labels. However, explicitly predicting future images or videos may include control-irrelevant appearance, while guidance derived from holistic future visual representations and shared g...
This work proposes FOCAL-VLA, a framework that combines subtask-guided geometry distillation with implicit world modeling to learn representations of current spatial structure and future interaction dynamics, and shows that FOCAL-VLA outperforms baselines on both simulation benchmarks and real-world manipulation tasks.
Zhi-Yuan Gao, Di Wen, Yan Zhan et al.· arXiv.org· 1 citation
IMLE-VLA is introduced, which replaces the iterative action head with a single-step conditional generator trained via conditional Implicit Maximum Likelihood Estimation (cIMLE), and promotes multimodal action coverage, avoiding the mode collapse of naive regression heads while eliminating multi-step sampling entirely.
Kian Hosseinkhani, Qin-He Peng, George Shramko et al.· arXiv.org· 2 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 9, 2026