Adversarial Attack with Pattern Replacement

TL;DR

This paper introduces a generative adversarial attack method that replaces input patterns with class-specific patterns to fool CNNs, demonstrated on MNIST.

Contribution

It presents a novel generative model for adversarial attacks that uses pattern replacement to deceive neural networks.

Findings

Effective attack on CNNs using pattern replacement

Generates subtle, class-specific adversarial patterns

Demonstrated on MNIST dataset

Abstract

We propose a generative model for adversarial attack. The model generates subtle but predictive patterns from the input. To perform an attack, it replaces the patterns of the input with those generated based on examples from some other class. We demonstrate our model by attacking CNN on MNIST.