2 papers
cs.RO2026
Selective Cross-View Consistency for World Action Models: Held-Out Viewpoint Robustness Without Test-Time Camera Information
Bingqi Huang, Bingchuan Wei, Yingkai Cai +1
World action models (WAMs) jointly denoise future video frames and robot actions, and the video prior is expected to generalize their control. Camera viewpoint change remains one o…
cs.RO2026
Cross-View Action Consistency for Camera-Robust Vision-Language-Action Policies
Bingqi Huang, Bingchuan Wei, Xuan Wang +2
Vision-language-action (VLA) policies fine-tuned from a fixed scene camera can fail when the camera is moved, even when the task, objects, language, and robot state are unchanged.…