Haiyang Zhang

From Forgetting to Robustness: Robust Class-Incremental Learning with CLIP

Text-Guided Attention is All You Need for Zero-Shot Robustness in Vision-Language Models