CV-DCLR: Causal-Visual Dynamic Label Refinement for Robust Zero-Shot Learning figure
AlphaXiv 中文概览(可滚动查看)