Adversarial Attacks on Variational Autoencoders
Computer Vision and Pattern Recognition
2018-06-13 v1 Machine Learning
Neural and Evolutionary Computing
Abstract
Adversarial attacks are malicious inputs that derail machine-learning models. We propose a scheme to attack autoencoders, as well as a quantitative evaluation framework that correlates well with the qualitative assessment of the attacks. We assess --- with statistically validated experiments --- the resistance to attacks of three variational autoencoders (simple, convolutional, and DRAW) in three datasets (MNIST, SVHN, CelebA), showing that both DRAW's recurrence and attention mechanism lead to better resistance. As autoencoders are proposed for compressing data --- a scenario in which their safety is paramount --- we expect more attention will be given to adversarial attacks on them.
Cite
@article{arxiv.1806.04646,
title = {Adversarial Attacks on Variational Autoencoders},
author = {George Gondim-Ribeiro and Pedro Tabacof and Eduardo Valle},
journal= {arXiv preprint arXiv:1806.04646},
year = {2018}
}