Transcending Adversarial Perturbations: Manifold-Aided Adversarial Examples with Legitimate Semantics

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, Shuai, Jiang, Xiaoyu, Ma, Xiaoguang
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917582473265152
author Li, Shuai
Jiang, Xiaoyu
Ma, Xiaoguang
author_facet Li, Shuai
Jiang, Xiaoyu
Ma, Xiaoguang
contents Deep neural networks were significantly vulnerable to adversarial examples manipulated by malicious tiny perturbations. Although most conventional adversarial attacks ensured the visual imperceptibility between adversarial examples and corresponding raw images by minimizing their geometric distance, these constraints on geometric distance led to limited attack transferability, inferior visual quality, and human-imperceptible interpretability. In this paper, we proposed a supervised semantic-transformation generative model to generate adversarial examples with real and legitimate semantics, wherein an unrestricted adversarial manifold containing continuous semantic variations was constructed for the first time to realize a legitimate transition from non-adversarial examples to adversarial ones. Comprehensive experiments on MNIST and industrial defect datasets showed that our adversarial examples not only exhibited better visual quality but also achieved superior attack transferability and more effective explanations for model vulnerabilities, indicating their great potential as generic adversarial examples. The code and pre-trained models were available at https://github.com/shuaili1027/MAELS.git.
format Preprint
id arxiv_https___arxiv_org_abs_2402_03095
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Transcending Adversarial Perturbations: Manifold-Aided Adversarial Examples with Legitimate Semantics
Li, Shuai
Jiang, Xiaoyu
Ma, Xiaoguang
Computer Vision and Pattern Recognition
Cryptography and Security
Machine Learning
Deep neural networks were significantly vulnerable to adversarial examples manipulated by malicious tiny perturbations. Although most conventional adversarial attacks ensured the visual imperceptibility between adversarial examples and corresponding raw images by minimizing their geometric distance, these constraints on geometric distance led to limited attack transferability, inferior visual quality, and human-imperceptible interpretability. In this paper, we proposed a supervised semantic-transformation generative model to generate adversarial examples with real and legitimate semantics, wherein an unrestricted adversarial manifold containing continuous semantic variations was constructed for the first time to realize a legitimate transition from non-adversarial examples to adversarial ones. Comprehensive experiments on MNIST and industrial defect datasets showed that our adversarial examples not only exhibited better visual quality but also achieved superior attack transferability and more effective explanations for model vulnerabilities, indicating their great potential as generic adversarial examples. The code and pre-trained models were available at https://github.com/shuaili1027/MAELS.git.
title Transcending Adversarial Perturbations: Manifold-Aided Adversarial Examples with Legitimate Semantics
topic Computer Vision and Pattern Recognition
Cryptography and Security
Machine Learning
url https://arxiv.org/abs/2402.03095