Paper ID: 2211.04973

Accelerating Adversarial Perturbation by 50% with Semi-backward Propagation

Zhiqi Bu

Adversarial perturbation plays a significant role in the field of adversarial robustness, which solves a maximization problem over the input data. We show that the backward propagation of such optimization can accelerate $2\times$ (and thus the overall optimization including the forward propagation can accelerate $1.5\times$), without any utility drop, if we only compute the output gradient but not the parameter gradient during the backward propagation.

Submitted: Nov 9, 2022