D-square-B: Deep Distribution Bound for Natural-looking Adversarial Attack
Preprint
Abstract
We propose a novel technique that can generate natural-looking adversarial examples by bounding the variations induced for internal activation values in some deep layer(s), through a distribution quantile bound and a polynomial barrier loss function. By bounding model internals instead of individual pixels, our attack admits perturbations closely coupled with the existing features of the original input, allowing the generated examples to be natural-looking while having diverse and often substantial pixel distances from the original input. Enforcing per-neuron distribution quantile bounds allows addressing the non-uniformity of internal activation values. Our evaluation on ImageNet and five different model architecture demonstrates that our attack is quite effective. Compared to the state-of-the-art pixel space attack, semantic attack, and feature space attack, our attack can achieve the same attack success/confidence level while having much more natural-looking adversarial perturbations. These perturbations piggy-back on existing local features and do not have any fixed pixel bounds.Keywords
All Related Versions
This publication has 9 references indexed in Scilit:
- Towards Feature Space Adversarial Attack by Style PerturbationProceedings of the AAAI Conference on Artificial Intelligence, 2021
- Sparse and Imperceivable Adversarial Attacks2019
- A Discussion of 'Adversarial Examples Are Not Bugs, They Are Features'Distill, 2019
- Harnessing the Vulnerability of Latent Layers in Adversarially Trained Models2019
- Neural Style Transfer: A ReviewIEEE Transactions on Visualization and Computer Graphics, 2019
- Disentangling Adversarial Robustness and Generalization2019
- Feature Space Perturbations Yield More Transferable Adversarial Examples2019
- Adversarial Examples Detection in Features Distance Spaces2019
- Deep Residual Learning for Image Recognition2016