ME-Net: Towards Effective Adversarial Robustness with Matrix Estimation

TL;DR

ME-Net introduces a novel defense mechanism against adversarial attacks by preprocessing images with pixel dropping and matrix estimation, effectively destroying adversarial noise while preserving essential image structure, leading to improved robustness.

Contribution

The paper presents ME-Net, a new adversarial defense method using matrix estimation that enhances robustness by reconstructing images after random pixel dropping.

Findings

ME-Net outperforms existing defenses on MNIST, CIFAR-10, SVHN, and Tiny-ImageNet.

The method effectively destroys adversarial noise while maintaining image integrity.

ME-Net improves robustness against both black-box and white-box attacks.

Abstract

Deep neural networks are vulnerable to adversarial attacks. The literature is rich with algorithms that can easily craft successful adversarial examples. In contrast, the performance of defense techniques still lags behind. This paper proposes ME-Net, a defense method that leverages matrix estimation (ME). In ME-Net, images are preprocessed using two steps: first pixels are randomly dropped from the image; then, the image is reconstructed using ME. We show that this process destroys the adversarial structure of the noise, while re-enforcing the global structure in the original image. Since humans typically rely on such global structures in classifying images, the process makes the network mode compatible with human perception. We conduct comprehensive experiments on prevailing benchmarks such as MNIST, CIFAR-10, SVHN, and Tiny-ImageNet. Comparing ME-Net with state-of-the-art defense mechanisms shows that ME-Net consistently outperforms prior techniques, improving robustness against both black-box and white-box attacks.