Improving Robustness of Deep Convolutional Neural Networks via Multiresolution Learning

TL;DR

This paper proposes multiresolution learning to enhance the robustness of deep convolutional neural networks against noise and adversarial attacks, showing improvements without sacrificing standard accuracy.

Contribution

It introduces multiresolution learning as a novel approach to improve DNN robustness across signal and image tasks, challenging the trade-off between accuracy and robustness.

Findings

Significant robustness improvements against noise and adversarial attacks.

Effective for small training datasets.

Maintains standard accuracy while enhancing robustness.

Abstract

The current learning process of deep learning, regardless of any deep neural network (DNN) architecture and/or learning algorithm used, is essentially a single resolution training. We explore multiresolution learning and show that multiresolution learning can significantly improve robustness of DNN models for both 1D signal and 2D signal (image) prediction problems. We demonstrate this improvement in terms of both noise and adversarial robustness as well as with small training dataset size. Our results also suggest that it may not be necessary to trade standard accuracy for robustness with multiresolution learning, which is, interestingly, contrary to the observation obtained from the traditional single resolution learning setting.