Over-the-Air Adversarial Attacks and Detection for Automatic Speaker Verification
- 12 January 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Audio, Speech and Language Processing
- Vol. 34, 1272-1285
Abstract
Automatic Speaker Verification (ASV) systems are widely used for voice-based identity verification, but recent studies have shown that they are vulnerable to both over-the-line (OTL) and over-the-air (OTA) adversarial attacks. Although various detection methods have been proposed, their robustness against diverse OTA attack scenarios has not been systematically assessed. In this work, we present AdvSV 2.0, a large-scale, open-source dataset for adversarial attacks on ASV systems. AdvSV 2.0 contains 628k utterances (800 hours) spanning 8 ASV models, 4 attack methods, and 9 physical playback-recording device combinations, and includes both OTL and OTA adversarial examples. To better connect digital adversarial optimization with the physical OTA channel, we further propose an OTA attack generation method based on a Neural Replay Simulator (NRS), which simulates replay-channel distortions and improves OTA attack success rates by up to 33.4% after air transmission. To defend against these attacks, we propose CODA-OCC, a COntrastive Domain-Aligned One-Class Classifier. Built on a one-class classification paradigm, CODA-OCC is trained only on bona fide speech and does not rely on any white-box assumption about the ASV backend. Experimental results on AdvSV 2.0 show that CODA-OCC achieves an EER of 11.2% and an AUC of 0.95, outperforming the purification-based baseline by 8.6% absolute EER, and providing a robust foundation for studying OTA adversarial robustness in ASV.Keywords
Funding Information
- National Natural Science Foundation of China (62376237)
- Program for Guangdong Introducing Innovative and Entrepreneurial Teams (2023ZT10X044)
- Korean Dermatological Association (ZDSYS20230626091302006)
- 2023 Shenzhen Stability Science Program
- Internal Project Fund from Shenzhen Research Institute of Big Data (T00120230002)
This publication has 38 references indexed in Scilit:
- Adversarial Sample Detection for Speaker Verification by Neural Vocoders2022
- Effective and Inconspicuous Over-the-Air Adversarial Examples with Adaptive Filtering2022
- Speaker recognition based on deep learning: An overviewNeural Networks, 2021
- FoolHD: Fooling Speaker Identification by Highly Imperceptible Adversarial Disturbances2021
- The ins and outs of speaker recognition: lessons from VoxSRC 20202021
- Libri-Light: A Benchmark for ASR with Limited or No Supervision2020
- Parallel Wavegan: A Fast Waveform Generation Model Based on Generative Adversarial Networks with Multi-Resolution Spectrogram2020
- Real-Time, Universal, and Robust Adversarial Attacks Against Speaker Recognition Systems2020
- Joint Decision of Anti-Spoofing and Automatic Speaker Verification by Multi-Task Learning With Contrastive LossIEEE Access, 2020
- X-Vectors: Robust DNN Embeddings for Speaker Recognition2018