Jia-Qi Lin

Learning Well-Structured Logits: Leveraging Vision–Language Complementarity for Open-World Test-Time Adaptation