LMD: A Learnable Mask Network to Detect Adversarial Examples for Speaker Verification

TL;DR

This paper introduces LMD, an attacker-independent, interpretable neural network-based method that detects adversarial examples in speaker verification by analyzing score variations caused by masked spectrograms.

Contribution

It proposes a novel, interpretable, attacker-independent detection method using score variation and masked spectrograms, trained solely on genuine examples.

Findings

Outperforms five state-of-the-art baselines in detection accuracy.

Effective against 12 different adversarial attack methods.

Provides a benchmark for detection-based ASV defenses.

Abstract

Although the security of automatic speaker verification (ASV) is seriously threatened by recently emerged adversarial attacks, there have been some countermeasures to alleviate the threat. However, many defense approaches not only require the prior knowledge of the attackers but also possess weak interpretability. To address this issue, in this paper, we propose an attacker-independent and interpretable method, named learnable mask detector (LMD), to separate adversarial examples from the genuine ones. It utilizes score variation as an indicator to detect adversarial examples, where the score variation is the absolute discrepancy between the ASV scores of an original audio recording and its transformed audio synthesized from its masked complex spectrogram. A core component of the score variation detector is to generate the masked spectrogram by a neural network. The neural network needs only genuine examples for training, which makes it an attacker-independent approach. Its interpretability lies that the neural network is trained to minimize the score variation of the targeted ASV, and maximize the number of the masked spectrogram bins of the genuine training examples. Its foundation is based on the observation that, masking out the vast majority of the spectrogram bins with little speaker information will inevitably introduce a large score variation to the adversarial example, and a small score variation to the genuine example. Experimental results with 12 attackers and two representative ASV systems show that our proposed method outperforms five state-of-the-art baselines. The extensive experimental results can also be a benchmark for the detection-based ASV defenses.