Enhancing Adversarial Transferability in Visual-Language Pre-training Models via Local Shuffle and Sample-based Attack.

Xin Liu, Aoyang Zhou, Kun He

Browse the full NAACL paper archive.