A Neural Rejection System Against Universal Adversarial Perturbations in Radio Signal Classification

TL;DR

This paper proposes a neural rejection system to defend against universal adversarial perturbations in radio signal classification, significantly improving robustness of deep learning models against such attacks.

Contribution

The paper introduces a novel neural rejection defense mechanism specifically designed to counter universal adversarial perturbations in radio signal classification tasks.

Findings

Neural rejection system significantly improves defense against universal adversarial perturbations.

The proposed method outperforms undefended neural networks in robustness.

White-box attack evaluations demonstrate the effectiveness of the defense.

Abstract

Advantages of deep learning over traditional methods have been demonstrated for radio signal classification in the recent years. However, various researchers have discovered that even a small but intentional feature perturbation known as adversarial examples can significantly deteriorate the performance of the deep learning based radio signal classification. Among various kinds of adversarial examples, universal adversarial perturbation has gained considerable attention due to its feature of being data independent, hence as a practical strategy to fool the radio signal classification with a high success rate. Therefore, in this paper, we investigate a defense system called neural rejection system to propose against universal adversarial perturbations, and evaluate its performance by generating white-box universal adversarial perturbations. We show that the proposed neural rejection system is able to defend universal adversarial perturbations with significantly higher accuracy than the undefended deep neural network.