Zobrazeno 1 - 1
of 1
pro vyhledávání: '"Xia, Sangzhou"'
Understanding the vulnerability of large-scale pre-trained vision-language models like CLIP against adversarial attacks is key to ensuring zero-shot generalization capacity on various downstream tasks. State-of-the-art defense mechanisms generally ad
Externí odkaz:
http://arxiv.org/abs/2405.11154
