mlsanityData to 14 Oct 2026

Certification of Speaker Recognition Models to Additive Perturbations

arXiv 2404.18791 ↗First posted 29 Apr 2024cs.SD cs.AI eess.AScited by 6

Not on the map. Why?

Abstract (arXiv)

Speaker recognition technology is applied to various tasks, from personal virtual assistants to secure access systems. However, the robustness of these systems against adversarial attacks, particularly to additive perturbations, remains a significant challenge. In this paper, we pioneer applying robustness certification techniques to speaker recognition, initially developed for the image domain. Our work covers this gap by transferring and improving randomized smoothing certification techniques against norm-bounded additive perturbations for classification and few-shot learning tasks to speaker recognition. We demonstrate the effectiveness of these methods on VoxCeleb 1 and 2 datasets for several models. We expect this work to improve the robustness of voice biometrics and accelerate the research of certification methods in the audio domain.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Builds on 34 of its 49 references that are in the map. First posted on arXiv 29 Apr 2024.

Builds on

34 of 49 references.

References · 34

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  2. 2013Intriguing properties of neural networkscited by 5,155
  3. 2019Certified Adversarial Robustness via Randomized Smoothingcited by 1,340
  4. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
  5. 2018VoxCeleb2: Deep Speaker Recognitioncited by 875
  6. 2017In Defense of the Triplet Loss for Person Re-Identificationcited by 786
  7. 2017VoxCeleb: a large-scale speaker identification datasetcited by 732
  8. 2021SpeechBrain: A General-Purpose Speech Toolkitcited by 406
  9. 2020Understanding the Behaviour of Contrastive Losscited by 398
  10. 2021MagFace: A Universal Representation for Face Recognition and Quality Assessmentcited by 334
  11. 2019Provably Robust Deep Learning via Adversarially Trained Smoothed Classifierscited by 330
  12. 2017Generalized End-to-End Loss for Speaker Verificationcited by 293
22 more
  1. 2020Understanding and Improving Fast Adversarial Trainingcited by 214
  2. 2019AdvHat: Real-world adversarial attack on ArcFace Face ID systemcited by 163
  3. 2021ASVspoof 2021: accelerating progress in spoofed and deepfake speech detectioncited by 161
  4. 2020Randomized Smoothing of All Shapes and Sizescited by 147
  5. 2020SoK: Certified Robustness for Deep Neural Networkscited by 105
  6. 2023CAM++: A Fast and Efficient Network for Speaker Verification Using Context-Aware Maskingcited by 102
  7. 2017Maximum Resilience of Artificial Neural Networkscited by 97
  8. 2022Wespeaker: A Research and Production oriented Speaker Embedding Learning Toolkitcited by 97
  9. 2020An Analysis of Adversarial Attacks and Defenses on Autonomous Driving Modelscited by 68
  10. 2021Meta Gradient Adversarial Attackcited by 55
  11. 2020TSS: Transformation-Specific Smoothing for Robustness Certificationcited by 49
  12. 2023An Enhanced Res2Net with Local and Global Feature Fusion for Speaker Verificationcited by 42
  13. 2020Second-Order Provable Defenses against Adversarial Attackscited by 40
  14. 2019Fastened CROWN: Tightened Neural Network Robustness Certificatescited by 39
  15. 2023OpenVoice: Versatile Instant Voice Cloningcited by 32
  16. 2021Improving the Adversarial Robustness for Speaker Verification by Self-Supervised Learningcited by 16
  17. 2020Extensions and limitations of randomized smoothing for robustness guaranteescited by 15
  18. 2019Real-world adversarial attack on MTCNN face detection systemcited by 14
  19. 2021Sequential Randomized Smoothing for Adversarially Robust Speech Recognitioncited by 13
  20. 2021Certified Robustness via Randomized Smoothing over Multiplicative Parameters of Input Transformationscited by 6
  21. 2022Smoothed Embeddings for Certified Few-Shot Learningcited by 6
  22. 2023On the Robustness of Split Learning against Adversarial Attackscited by 6

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 4; 2026: 2.

6 papers.

  1. 2025Position: Certified Robustness Does Not (Yet) Imply Model Securitycited by 4
  2. 2025Optimal Hyperspectral Undersampling Strategy for Satellite Imagingcited by 0
  3. 2025Novel Loss-Enhanced Universal Adversarial Patches for Sustainable Speaker Privacycited by 0
  4. 2025Empirical evaluation of the Frank-Wolfe methods for constructing white-box adversarial attackscited by 0
  5. 2026Contract And Conquer: How to Provably Compute Adversarial Examples for a Black-Box Model?cited by 0
  6. 2026Representation Matters in Randomized Smoothing for Audio Classificationcited by 0