Jiawei Li

~Jiawei_Li4

One Perturbation is Enough: On Generating Universal Adversarial Perturbations against Vision-Language Pre-training Models

Efficient Differentiable Approximation of the Generalized Low-rank Regularization