Adversarial robustness of VAEs through the lens of local geometry

Asif Khan, Amos Storkey

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

In an unsupervised attack on variational autoencoders (VAEs), an adversary finds a small perturbation in an input sample that significantly changes its latent space encoding, thereby compromising the reconstruction for a fixed decoder. A known reason for such vulnerability is the distortions in the latent space resulting from a mismatch between approximated latent posterior and a prior distribution. Consequently, a slight change in an input sample can move its encoding to a low/zero density region in the latent space resulting in an unconstrained generation. This paper demonstrates that an optimal way for an adversary to attack VAEs is to exploit a directional bias of a stochastic pullback metric tensor induced by the encoder and decoder networks. The pullback metric tensor of an encoder measures the change in infinitesimal latent volume from an input to a latent space. Thus, it can be viewed as a lens to analyse the effect of input perturbations leading to latent space distortions. We propose robustness evaluation scores using the eigenspectrum of a pullback metric tensor. Moreover, we empirically show that the scores correlate with the robustness parameter β of the β-VAE. Since increasing β also degrades reconstruction quality, we demonstrate a simple alternative using mixup training to fill the empty regions in the latent space, thus improving robustness with improved reconstruction.
Original languageEnglish
Title of host publicationProceedings of The 26th International Conference on Artificial Intelligence and Statistics
EditorsFrancisco Ruiz, Jennifer Dy, Jan-Willem van de Meent
Number of pages14
Publication statusPublished - 25 Apr 2023

Publication series

NameProceedings of Machine Learning Research
ISSN (Electronic)2640-3498


Dive into the research topics of 'Adversarial robustness of VAEs through the lens of local geometry'. Together they form a unique fingerprint.

Cite this