Semantic Preserving Adversarial Attack Generation with Autoencoder and Genetic Algorithm

2022/08/25 by Xinyi Wang, Wang, Xinyi, Simon Yusuf Enoch +3 · 1 citation
Computer Science · #Adversarial Robustness in Machine Learning #Anomaly Detection Techniques and Applications #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Neural and Evolutionary Computing (cs.NE)

paper · pdf · doi:10.48550/arxiv.2208.12230

openalex publication_date 2022/08/25 · openalex created_date 2022/08/27 · openalex updated_date 2026/07/28

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.

Cited by

• Beyond Vulnerabilities: A Survey of Adversarial Attacks as Both Threats and Defenses in Computer Vision Systems

Related

• Robust Deep Learning Models Against Semantic-Preserving Adversarial Attack
• Semantic Autoencoder and Its Potential Usage for Adversarial Attack
• Defense Against Adversarial Attacks using Convolutional Auto-Encoders
• A Generative Model based Adversarial Security of Deep Learning and\n Linear Classifier Models
• Combatting Adversarial Attacks through Denoising and Dimensionality\n Reduction: A Cascaded Autoencoder Approach