MMODELYST
Papers/Fast Adversarial Attacks with Gradient Prediction
PAP

Fast Adversarial Attacks with Gradient Prediction

May 14, 2026

arXiv
Abstract

Generating adversarial examples at scale is a core primitive for robustness evaluation, adversarial training, and red-teaming, yet even "fast" attacks such as FGSM remain throughput-limited by the cost of a backward pass. We introduce a family of attacks that eliminates the backward pass by predicting the input gradient from forward-pass hidden states via a lightweight linear regression. The approach is motivated by a kernel view of neural networks and is exact in the Neural Tangent Kernel regime, while remaining effective for practical finite-width models. Empirically, our methods recover much of FGSM's attack performance while using only a small fraction of the time, corresponding to a $532\%$ increase in throughput. These results suggest gradient prediction as a simple and general route to significantly faster adversarial generation under realistic wall-clock constraints.

Select text to highlight · click a highlight to remove · saved in this browser only
Authors
Kamil Ciosek, Aleksandr V. Petrov, Nicolò Felicioni, Konstantina Palla
Your notes (browser-local)
saved
arXiv:2605.14868