English

Memory Defense: More Robust Classification via a Memory-Masking Autoencoder

Computer Vision and Pattern Recognition 2022-02-08 v1 Machine Learning

Abstract

Many deep neural networks are susceptible to minute perturbations of images that have been carefully crafted to cause misclassification. Ideally, a robust classifier would be immune to small variations in input images, and a number of defensive approaches have been created as a result. One method would be to discern a latent representation which could ignore small changes to the input. However, typical autoencoders easily mingle inter-class latent representations when there are strong similarities between classes, making it harder for a decoder to accurately project the image back to the original high-dimensional space. We propose a novel framework, Memory Defense, an augmented classifier with a memory-masking autoencoder to counter this challenge. By masking other classes, the autoencoder learns class-specific independent latent representations. We test the model's robustness against four widely used attacks. Experiments on the Fashion-MNIST & CIFAR-10 datasets demonstrate the superiority of our model. We make available our source code at GitHub repository: https://github.com/eashanadhikarla/MemDefense

Keywords

Cite

@article{arxiv.2202.02595,
  title  = {Memory Defense: More Robust Classification via a Memory-Masking Autoencoder},
  author = {Eashan Adhikarla and Dan Luo and Brian D. Davison},
  journal= {arXiv preprint arXiv:2202.02595},
  year   = {2022}
}

Comments

11 pages

R2 v1 2026-06-24T09:21:50.731Z