Sparse2Act: Learning Action-Aligned Sparse 3D Representations for Cross-Domain Robot Manipulation figure
AlphaXiv 中文概览(可滚动查看)