# A principled approach for generating adversarial images under non-smooth dissimilarity metrics

Source: [A principled approach for generating adversarial images under non-smooth dissimilarity metrics](<https://www.sota2.com/research/paper/68fa66781e209131cea1059f>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T21:50:28.993Z

Authors: Aram-Alexandre Pooladian, Chris Finlay, Tim Hoheisel, Adam Oberman

Published: 2019-08-05T14:57:01.000Z

## Abstract

Deep neural networks perform well on real world data but are prone to adversarial perturbations: small changes in the input easily lead to misclassification. In this work, we propose an attack methodology not only for cases where the perturbations are measured by $\\ell\_p$ norms, but in fact any adversarial dissimilarity metric with a closed proximal form. This includes, but is not limited to, $\\ell\_1, \\ell\_2$, and $\\ell\_\\infty$ perturbations; the $\\ell\_0$ counting "norm" (i.e. true sparseness); and the total variation seminorm, which is a (non-$\\ell\_p$) convolutional dissimilarity measuring local pixel changes. Our approach is a natural extension of a recent adversarial attack method, and eliminates the differentiability requirement of the metric. We demonstrate our algorithm, ProxLogBarrier, on the MNIST, CIFAR10, and ImageNet-1k datasets. We consider undefended and defended models, and show that our algorithm easily transfers to various datasets. We observe that ProxLogBarrier outperforms a host of modern adversarial attacks specialized for the $\\ell\_0$ case. Moreover, by altering images in the total variation seminorm, we shed light on a new class of perturbations that exploit neighboring pixel information.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1908.01667>)

[PDF](<https://arxiv.org/pdf/1908.01667.pdf>)
