Black-box Evolutionary Search for Adversarial Examples against Deep Image Classifiers in Non-Targeted Attacks

0534345 - ÚI 2021 RIV US eng C - Konferenční příspěvek (zahraniční konf.)
Procházka, Štěpán - Neruda, Roman
Black-box Evolutionary Search for Adversarial Examples against Deep Image Classifiers in Non-Targeted Attacks.
2020 International Joint Conference on Neural Networks (IJCNN): Conference Proceedings. Piscataway: IEEE, 2020, č. článku 9207688. ISBN 978-1-7281-6926-2.
[IJCNN 2020: International Joint Conference on Neural Networks /33./. Glasgow (GB), 19.07.2020-24.07.2020]
Grant CEP: GA ČR(CZ) GA18-23827S
Institucionální podpora: RVO:67985807
Klíčová slova: Machine learning * Perturbation methods * Task analysis * Data models * Computer architecture * Sociology * Statistics
Obor OECD: Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Machine learning models exhibit vulnerability to adversarial examples i.e., artificially created inputs that become misinterpreted. The goal of this paper is to explore non-targeted black-box adversarial attacks on deep networks performing image classification. The original evolutionary algorithm for generating adversarial examples is proposed that employs a guided multi-objective search through the space of perturbed images. The efficiency of attacks is validated by experiments with the CIFAR-10 data set. The experimental results verify the usability of our approach against deep convolutional neural networks.
Trvalý link: http://hdl.handle.net/11104/0312558