English

Multi-Target Backdoor Attacks Against Speaker Recognition

Sound 2025-10-10 v3 Machine Learning

Abstract

In this work, we propose a multi-target backdoor attack against speaker identification using position-independent clicking sounds as triggers. Unlike previous single-target approaches, our method targets up to 50 speakers simultaneously, achieving success rates of up to 95.04%. To simulate more realistic attack conditions, we vary the signal-to-noise ratio between speech and trigger, demonstrating a trade-off between stealth and effectiveness. We further extend the attack to the speaker verification task by selecting the most similar training speaker - based on cosine similarity - as a proxy target. The attack is most effective when target and enrolled speaker pairs are highly similar, reaching success rates of up to 90% in such cases.

Keywords

Cite

@article{arxiv.2508.08559,
  title  = {Multi-Target Backdoor Attacks Against Speaker Recognition},
  author = {Alexandrine Fortier and Sonal Joshi and Thomas Thebaud and Jesús Villalba and Najim Dehak and Patrick Cardinal},
  journal= {arXiv preprint arXiv:2508.08559},
  year   = {2025}
}

Comments

Accepted to IEEE Automatic Speech Recognition and Understanding Workshop 2025

R2 v1 2026-07-01T04:45:25.791Z