Low-Rank Adversarial PGD Attack

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Savostianova, Dayana, Zangrando, Emanuele, Tudisco, Francesco
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910652901097472
author Savostianova, Dayana
Zangrando, Emanuele
Tudisco, Francesco
author_facet Savostianova, Dayana
Zangrando, Emanuele
Tudisco, Francesco
contents Adversarial attacks on deep neural network models have seen rapid development and are extensively used to study the stability of these networks. Among various adversarial strategies, Projected Gradient Descent (PGD) is a widely adopted method in computer vision due to its effectiveness and quick implementation, making it suitable for adversarial training. In this work, we observe that in many cases, the perturbations computed using PGD predominantly affect only a portion of the singular value spectrum of the original image, suggesting that these perturbations are approximately low-rank. Motivated by this observation, we propose a variation of PGD that efficiently computes a low-rank attack. We extensively validate our method on a range of standard models as well as robust models that have undergone adversarial training. Our analysis indicates that the proposed low-rank PGD can be effectively used in adversarial training due to its straightforward and fast implementation coupled with competitive performance. Notably, we find that low-rank PGD often performs comparably to, and sometimes even outperforms, the traditional full-rank PGD attack, while using significantly less memory.
format Preprint
id arxiv_https___arxiv_org_abs_2410_12607
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Low-Rank Adversarial PGD Attack
Savostianova, Dayana
Zangrando, Emanuele
Tudisco, Francesco
Machine Learning
Artificial Intelligence
Numerical Analysis
Adversarial attacks on deep neural network models have seen rapid development and are extensively used to study the stability of these networks. Among various adversarial strategies, Projected Gradient Descent (PGD) is a widely adopted method in computer vision due to its effectiveness and quick implementation, making it suitable for adversarial training. In this work, we observe that in many cases, the perturbations computed using PGD predominantly affect only a portion of the singular value spectrum of the original image, suggesting that these perturbations are approximately low-rank. Motivated by this observation, we propose a variation of PGD that efficiently computes a low-rank attack. We extensively validate our method on a range of standard models as well as robust models that have undergone adversarial training. Our analysis indicates that the proposed low-rank PGD can be effectively used in adversarial training due to its straightforward and fast implementation coupled with competitive performance. Notably, we find that low-rank PGD often performs comparably to, and sometimes even outperforms, the traditional full-rank PGD attack, while using significantly less memory.
title Low-Rank Adversarial PGD Attack
topic Machine Learning
Artificial Intelligence
Numerical Analysis
url https://arxiv.org/abs/2410.12607