English

Semantic Preserving Adversarial Attack Generation with Autoencoder and Genetic Algorithm

Machine Learning 2022-08-26 v1 Cryptography and Security Neural and Evolutionary Computing

Abstract

Widely used deep learning models are found to have poor robustness. Little noises can fool state-of-the-art models into making incorrect predictions. While there is a great deal of high-performance attack generation methods, most of them directly add perturbations to original data and measure them using L_p norms; this can break the major structure of data, thus, creating invalid attacks. In this paper, we propose a black-box attack, which, instead of modifying original data, modifies latent features of data extracted by an autoencoder; then, we measure noises in semantic space to protect the semantics of data. We trained autoencoders on MNIST and CIFAR-10 datasets and found optimal adversarial perturbations using a genetic algorithm. Our approach achieved a 100% attack success rate on the first 100 data of MNIST and CIFAR-10 datasets with less perturbation than FGSM.

Keywords

Cite

@article{arxiv.2208.12230,
  title  = {Semantic Preserving Adversarial Attack Generation with Autoencoder and Genetic Algorithm},
  author = {Xinyi Wang and Simon Yusuf Enoch and Dong Seong Kim},
  journal= {arXiv preprint arXiv:2208.12230},
  year   = {2022}
}

Comments

8 pages conference paper, accepted for publication in IEEE GLOBECOM 2022

R2 v1 2026-06-25T01:58:56.182Z