# Constrained Gradient Descent: A Powerful and Principled Evasion Attack Against Neural Networks

Source: [Constrained Gradient Descent: A Powerful and Principled Evasion Attack Against Neural Networks](<https://www.sota2.com/research/paper/68fa6cd91e209131cea36414>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T00:19:37.765Z

Authors: Weiran Lin, Keane Lucas, Lujo Bauer, Michael K. Reiter, Mahmood Sharif

Published: 2021-12-28T17:36:58.000Z

## Abstract

We propose new, more efficient targeted white-box attacks against deep neural networks. Our attacks better align with the attacker's goal: (1) tricking a model to assign higher probability to the target class than to any other class, while (2) staying within an $\\epsilon$-distance of the attacked input. First, we demonstrate a loss function that explicitly encodes (1) and show that Auto-PGD finds more attacks with it. Second, we propose a new attack method, Constrained Gradient Descent (CGD), using a refinement of our loss function that captures both (1) and (2). CGD seeks to satisfy both attacker objectives -- misclassification and bounded $\\ell\_{p}$-norm -- in a principled manner, as part of the optimization, instead of via ad hoc post-processing techniques (e.g., projection or clipping). We show that CGD is more successful on CIFAR10 (0.9--4.2%) and ImageNet (8.6--13.6%) than state-of-the-art attacks while consuming less time (11.4--18.8%). Statistical tests confirm that our attack outperforms others against leading defenses on different datasets and values of $\\epsilon$.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2112.14232>)

[PDF](<https://arxiv.org/pdf/2112.14232.pdf>)
