Adversarial Attacks on GMM I-Vector Based Speaker Verification Systems

@article{Li2020AdversarialAO,
  title={Adversarial Attacks on GMM I-Vector Based Speaker Verification Systems},
  author={X. Li and Jinghua Zhong and Xixin Wu and J. Yu and Xunying Liu and Helen Meng},
  journal={ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  year={2020},
  pages={6579-6583}
}
  • X. Li, Jinghua Zhong, +3 authors Helen Meng
  • Published 2020
  • Computer Science, Engineering
  • ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
  • This work investigates the vulnerability of Gaussian Mixture Model (GMM) i-vector based speaker verification systems to adversarial attacks, and the transferability of adversarial samples crafted from GMM i-vector based systems to x-vector based systems. In detail, we formulate the GMM i-vector system as a scoring function of enrollment and testing utterance pairs. Then we leverage the fast gradient sign method (FGSM) to optimize testing utterances for adversarial samples generation. These… CONTINUE READING
    8 Citations

    Figures, Tables, and Topics from this paper.

    Adversarial Attack and Defense Strategies for Deep Speaker Recognition Systems
    • 2
    • PDF
    Investigating Robustness of Adversarial Samples Detection for Automatic Speaker Verification
    • 2
    • PDF
    The Attacker's Perspective on Automatic Speaker Verification: An Overview
    • 6
    • PDF
    Defending Your Voice: Adversarial Attack on Voice Conversion
    • 2
    • PDF
    Replay and Synthetic Speech Detection with Res2net Architecture

    References

    SHOWING 1-10 OF 29 REFERENCES
    Fooling End-To-End Speaker Verification With Adversarial Examples
    • 67
    • PDF
    Vulnerability of speaker verification systems against voice conversion spoofing attacks: The case of telephone speech
    • 147
    • PDF
    Are Image-Agnostic Universal Adversarial Perturbations for Face Recognition Difficult to Detect?
    • 23
    • PDF
    Adversarial Machine Learning at Scale
    • 1,141
    • PDF
    A study on replay attack and anti-spoofing for text-dependent speaker verification
    • 112
    • PDF
    Deep neural networks for small footprint text-dependent speaker verification
    • 516
    • PDF
    Front-End Factor Analysis For Speaker Verification
    • Florin Curelaru
    • Computer Science
    • 2018 International Conference on Communications (COMM)
    • 2018
    • 940
    Deep neural network-based speaker embeddings for end-to-end speaker verification
    • 229
    • PDF
    Audio Adversarial Examples: Targeted Attacks on Speech-to-Text
    • 417
    • PDF