LatentPoison - Adversarial Attacks On The Latent Space
This exposes a security risk in unsupervised generative models, which is incremental as it extends adversarial attack methods to latent spaces.
The paper tackles the problem of latent space vulnerability in deep variational autoencoders by demonstrating that adversarial perturbations can flip class predictions while keeping classification probabilities unchanged, making attacks undetectable to decoders.
Robustness and security of machine learning (ML) systems are intertwined, wherein a non-robust ML system (classifiers, regressors, etc.) can be subject to attacks using a wide variety of exploits. With the advent of scalable deep learning methodologies, a lot of emphasis has been put on the robustness of supervised, unsupervised and reinforcement learning algorithms. Here, we study the robustness of the latent space of a deep variational autoencoder (dVAE), an unsupervised generative framework, to show that it is indeed possible to perturb the latent space, flip the class predictions and keep the classification probability approximately equal before and after an attack. This means that an agent that looks at the outputs of a decoder would remain oblivious to an attack.