Researcher, University of Hong Kong
1 paper at NeurIPS 2025
Our approach improves Vision Transformer Dense Representations via Self-Distillation