Adversarial Attack and Defense Strategies for Deep Speaker Recognition Systems

TL;DR

This paper investigates the vulnerability of deep speaker recognition systems to adversarial attacks and evaluates defense strategies, finding adversarial training with PGD as the most effective method to enhance robustness.

Contribution

It provides a comprehensive analysis of adversarial attacks and defenses in speaker recognition, including experimental comparisons and ablation studies, which are limited in existing research.

Findings

Adversarial attacks can reduce speaker recognition accuracy from 94% to 0%.

Adversarial training with PGD is the most effective defense method.

Speaker recognition systems are highly vulnerable to adversarial perturbations.

Abstract

Robust speaker recognition, including in the presence of malicious attacks, is becoming increasingly important and essential, especially due to the proliferation of several smart speakers and personal agents that interact with an individual's voice commands to perform diverse, and even sensitive tasks. Adversarial attack is a recently revived domain which is shown to be effective in breaking deep neural network-based classifiers, specifically, by forcing them to change their posterior distribution by only perturbing the input samples by a very small amount. Although, significant progress in this realm has been made in the computer vision domain, advances within speaker recognition is still limited. The present expository paper considers several state-of-the-art adversarial attacks to a deep speaker recognition system, employing strong defense methods as countermeasures, and reporting on several ablation studies to obtain a comprehensive understanding of the problem. The experiments show that the speaker recognition systems are vulnerable to adversarial attacks, and the strongest attacks can reduce the accuracy of the system from 94% to even 0%. The study also compares the performances of the employed defense methods in detail, and finds adversarial training based on Projected Gradient Descent (PGD) to be the best defense method in our setting. We hope that the experiments presented in this paper provide baselines that can be useful for the research community interested in further studying adversarial robustness of speaker recognition systems.