VistaVLA: Geometry- and Semantic-Aware 3D Gaussian-Grounded VLA for Robotic Manipulation figure
AlphaXiv 中文概览(可滚动查看)