Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
97 tokens/sec
GPT-4o
53 tokens/sec
Gemini 2.5 Pro Pro
44 tokens/sec
o3 Pro
5 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Differentially Private Adversarial Auto-Encoder to Protect Gender in Voice Biometrics (2307.02135v1)

Published 5 Jul 2023 in eess.AS

Abstract: Over the last decade, the use of Automatic Speaker Verification (ASV) systems has become increasingly widespread in response to the growing need for secure and efficient identity verification methods. The voice data encompasses a wealth of personal information, which includes but is not limited to gender, age, health condition, stress levels, and geographical and socio-cultural origins. These attributes, known as soft biometrics, are private and the user may wish to keep them confidential. However, with the advancement of machine learning algorithms, soft biometrics can be inferred automatically, creating the potential for unauthorized use. As such, it is crucial to ensure the protection of these personal data that are inherent within the voice while retaining the utility of identity recognition. In this paper, we present an adversarial Auto-Encoder--based approach to hide gender-related information in speaker embeddings, while preserving their effectiveness for speaker verification. We use an adversarial procedure against a gender classifier and incorporate a layer based on the Laplace mechanism into the Auto-Encoder architecture. This layer adds Laplace noise for more robust gender concealment and ensures differential privacy guarantees during inference for the output speaker embeddings. Experiments conducted on the VoxCeleb dataset demonstrate that speaker verification tasks can be effectively carried out while concealing speaker gender and ensuring differential privacy guarantees; moreover, the intensity of the Laplace noise can be tuned to select the desired trade-off between privacy and utility.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (34)
  1. Deep Learning with Differential Privacy. In Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security (Vienna, Austria) (CCS ’16). Association for Computing Machinery, New York, NY, USA, 308–318. https://doi.org/10.1145/2976749.2978318
  2. John M Abowd. 2018. The US Census Bureau adopts differential privacy. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2867–2867.
  3. Emotionless: Privacy-preserving speech analysis for voice assistants. arXiv preprint arXiv:1908.03632 (2019).
  4. Beyond Voice Identity Conversion: Manipulating Voice Attributes by Adversarial Learning of Structured Disentangled Representations. arXiv preprint arXiv:2107.12346 (2021).
  5. Learning privacy-enhancing face representations through feature disentanglement. In 2020 15th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2020). IEEE, 495–502.
  6. J Michelle Brock and Ralph De Haas. 2021. EVIDENCE FROM BANKERS IN THE LAB. (2021).
  7. Privacy-Preserving Voice Anti-Spoofing Using Secure Multi-Party Computation. In Interspeech 2021. ISCA, Brno, Czech Republic, 856–860. https://doi.org/10.21437/Interspeech.2021-983
  8. VoxCeleb2: Deep Speaker Recognition. In INTERSPEECH.
  9. Anastasia Cozarenco and Ariane Szafarz. 2018. Gender biases in bank lending: Lessons from microcredit in France. Journal of Business Ethics 147 (2018), 631–650.
  10. ECAPA-TDNN Embeddings for Speaker Diarization. In Proc. Interspeech 2021. 3560–3564. https://doi.org/10.21437/Interspeech.2021-941
  11. Calibrating noise to sensitivity in private data analysis. In Theory of Cryptography: Third Theory of Cryptography Conference, TCC 2006, New York, NY, USA, March 4-7, 2006. Proceedings 3. Springer, 265–284.
  12. The algorithmic foundations of differential privacy. Found. Trends Theor. Comput. Sci. 9, 3-4 (2014), 211–407.
  13. What can we learn privately? SIAM J. Comput. 40, 3 (2011), 793–826.
  14. Certified robustness to adversarial examples with differential privacy. In 2019 IEEE Symposium on Security and Privacy (SP). IEEE, 656–672.
  15. Improving speaker recognition by biometric voice deconstruction. Frontiers in bioengineering and biotechnology 3 (2015), 126.
  16. Multi-IVE: Privacy Enhancement of Multiple Soft-Biometrics in Face Embeddings. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 323–331.
  17. Semi-adversarial networks: Convolutional autoencoders for imparting privacy to face images. In 2018 International Conference on Biometrics (ICB). IEEE, 82–89.
  18. Gender privacy: An ensemble of semi adversarial networks for confounding arbitrary gender classifiers. In 2018 IEEE 9th International Conference on Biometrics Theory, Applications and Systems (BTAS). IEEE, 1–10.
  19. Flowsan: Privacy-enhancing semi-adversarial networks to confound arbitrary face-based gender classifiers. IEEE Access 7 (2019), 99735–99745.
  20. SensitiveNets: Learning agnostic representations with application to face images. IEEE Transactions on Pattern Analysis and Machine Intelligence 43, 6 (2020), 2158–2164.
  21. Voxceleb: a large-scale speaker identification dataset. arXiv preprint arXiv:1706.08612 (2017).
  22. Karthik Nandakumar and Anil K. Jain. 2009. Soft Biometrics. Springer US, Boston, MA, 1235–1239. https://doi.org/10.1007/978-0-387-73003-5_225
  23. Smooth sensitivity and sampling in private data analysis. In Proceedings of the thirty-ninth annual ACM symposium on Theory of computing. 75–84.
  24. Adversarial disentanglement of speaker representation for attribute-driven privacy preservation. arXiv preprint arXiv:2012.04454 (2020).
  25. Speech2face: Learning the face behind a voice. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 7539–7548.
  26. Improving Transformer-based Networks With Locality For Automatic Speaker Verification. arXiv preprint arXiv:2302.08639 (2023).
  27. X-Vectors: Robust DNN Embeddings for Speaker Recognition. In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 5329–5333. https://doi.org/10.1109/ICASSP.2018.8461375
  28. Gender-adversarial networks for face privacy preserving. IEEE Internet of Things Journal 9, 18 (2022), 17568–17576.
  29. Suppressing gender and age in face templates using incremental variable elimination. In 2019 International Conference on Biometrics (ICB). IEEE, 1–8.
  30. PE-MIU: A training-free privacy-enhancing face recognition approach based on minimum information units. IEEE Access 8 (2020), 93635–93647.
  31. Privacy-Preserving Convolutional Neural Networks Using Homomorphic Encryption. In 2022 International Workshop on Biometrics and Forensics (IWBF). 1–6. https://doi.org/10.1109/IWBF55382.2022.9794535
  32. One Source to Detect them All: Gender, Age, and Emotion Detection from Voice. In 2021 IEEE 45th Annual Computers, Software, and Applications Conference (COMPSAC). 338–343. https://doi.org/10.1109/COMPSAC51774.2021.00055
  33. Bias and variance of post-processing in differential privacy. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 35. 11177–11184.
  34. Medical imaging deep learning with differential privacy. Scientific Reports 11, 1 (2021), 1–8.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (7)
  1. Michele Panariello (12 papers)
  2. Oualid Zari (3 papers)
  3. Ismet Kerenciler (1 paper)
  4. Imen Chihaoui (1 paper)
  5. Massimiliano Todisco (55 papers)
  6. Melek Önen (10 papers)
  7. Oubaïda Chouchane (1 paper)
Citations (3)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now