Universal Adversarial Perturbations for Vision-Language Pre-trained Models.

Peng-Fei Zhang, Zi Huang, Guangdong Bai

Browse the full SIGIR paper archive.