Adversarial examples for generative models

Abstract

We explore methods of producing adversarial examples on deep generative models such as the variational autoencoder (VAE) and the VAE-GAN. Deep learning architectures are known to be vulnerable to adversarial examples, but previous work has focused on the application of adversarial examples to classification tasks. Deep generative models have recently become popular due to their ability to model input data distributions and generate realistic examples from those distributions. We present three classes of attacks on the VAE and VAE-GAN architectures and demonstrate them against networks trained on MNIST, SVHN and CelebA. Our first attack leverages classification-based adversaries by attaching a classifier to the trained encoder of the target generative model, which can then be used to indirectly manipulate the latent representation. Our second attack directly uses the VAE loss function to generate a target reconstruction image from the adversarial example. Our third attack moves beyond relying on classification or the standard loss for the gradient and directly optimizes against differences in source and target latent representations. We also motivate why an attacker might be interested in deploying such techniques against a target generative network.

References in corpus (1)

• Deep Convolutional Inverse Graphics Network

Cited by in corpus (8)

• Towards Crafting Text Adversarial Samples
• CycleGAN, a Master of Steganography
• UPSET and ANGRI : Breaking High Performance Image Classifiers
• Blocking Transferability of Adversarial Examples in Black-Box Learning Systems
• Adversarial Example Defenses: Ensembles of Weak Defenses are not Strong
• Detecting Adversarial Attacks on Neural Network Policies with Visual Foresight
• LatentPoison - Adversarial Attacks On The Latent Space
• Adversarial Security Attacks and Perturbations on Machine Learning and Deep Learning Methods