Vision-language-action (VLA) policies must ground language instructions in the right objects, contact points, and target locations for reliable manipulation. Human gaze provides a natural signal for this grounding, but existing robot datasets rarely pair real human gaze with the multi-camera observations used by modern...
Anupam Pani, Wang-Gui Yuan, Qi-Fan Wang et al.· 0 citations
Modeling human-object interaction (HOI) in real environments requires jointly representing whole-body motion, manipulated objects, and scene geometry, yet existing datasets trade off scalable capture against structured, scene-aligned supervision. We introduce SituHOI, a large-scale RGB-D dataset containing 18,429 annot...
Qi-Fan Wang, Qi-Hang Fang, Bo-Ao Shi et al.· 0 citations
A vision-language-action (VLA) policy with a flow-matching action expert generates each action chunk (a short command sequence) by integrating a learned velocity field; once its weights are fixed, the success or failure of an earlier rollout cannot change the chunk generated now. Concurrent test-time methods give a fro...
Physical contact often determines how a humanoid should respond during loco-manipulation, yet vision and proprioception alone are often insufficient to characterize physical interaction, especially when the contact region is occluded. Unlike sparse force or torque measurements at predefined regions, distributed tactile...
Zi-Hao Wang, Shu-Tong Liu, Si-Qi Zheng et al.· 0 citations
SlackDrive is proposed, a pre-inference compute allocator that reuses realized latency to select the compute budget of each control step before model execution, complementing existing profiling and resource scheduling while preserving the driving backbone and its compute actuator.
Xiao-Huan Pei, Heng-Guang Zhou, Yuan-Hao Ban et al.· 0 citations
By reformulating MI estimation as an inference task, InfoAtlas establishes a foundation for real-time dependency analysis and generalizes effectively to complex, real-world scenarios.