Adversarial Filters for Secure Modulation Classification

TL;DR

This paper introduces novel adversarial filtering techniques to enhance secure modulation classification by maximizing accuracy at legitimate receivers while minimizing it at eavesdroppers, offering efficient and robust solutions.

Contribution

The paper proposes new adversarial filtering algorithms for secure modulation classification, improving over traditional perturbation methods with efficiency and robustness.

Findings

GAF and FGFM outperform additive adversarial perturbations

Algorithms are computationally efficient and power-efficient

Effective at low SNR conditions

Abstract

Modulation Classification (MC) refers to the problem of classifying the modulation class of a wireless signal. In the wireless communications pipeline, MC is the first operation performed on the received signal and is critical for reliable decoding. This paper considers the problem of secure modulation classification, where a transmitter (Alice) wants to maximize MC accuracy at a legitimate receiver (Bob) while minimizing MC accuracy at an eavesdropper (Eve). The contribution of this work is to design novel adversarial learning techniques for secure MC. In particular, we present adversarial filtering based algorithms for secure MC, in which Alice uses a carefully designed adversarial filter to mask the transmitted signal, that can maximize MC accuracy at Bob while minimizing MC accuracy at Eve. We present two filtering based algorithms, namely gradient ascent filter (GAF), and a fast gradient filter method (FGFM), with varying levels of complexity. Our proposed adversarial filtering based approaches significantly outperform additive adversarial perturbations (used in the traditional ML community and other prior works on secure MC) and also have several other desirable properties. In particular, GAF and FGFM algorithms are a) computational efficient (allow fast decoding at Bob), b) power-efficient (do not require excessive transmit power at Alice); and c) SNR efficient (i.e., perform well even at low SNR values at Bob).