mlsanityData to 14 Oct 2026

Privacy-Preserving Speech Representation Learning using Vector Quantization

arXiv 2203.09518 ↗First posted 15 Mar 2022eess.AS cs.AI cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

With the popularity of virtual assistants (e.g., Siri, Alexa), the use of speech recognition is now becoming more and more widespread.However, speech signals contain a lot of sensitive information, such as the speaker's identity, which raises privacy concerns.The presented experiments show that the representations extracted by the deep layers of speech recognition networks contain speaker information.This paper aims to produce an anonymous representation while preserving speech recognition performance.To this end, we propose to use vector quantization to constrain the representation space and induce the network to suppress the speaker identity.The choice of the quantization dictionary size allows to configure the trade-off between utility (speech recognition) and privacy (speaker identity concealment).

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 11 of its 22 references that are in the map. First posted on arXiv 15 Mar 2022.

Builds on

11 of 22 references.

References · 11

  1. 2017Neural Discrete Representation Learningcited by 4,712
  2. 2013Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computationcited by 2,079
  3. 2018Fast Decoding in Sequence Models using Discrete Latent Variablescited by 104
  4. 2019Unsupervised speech representation learning using WaveNet autoencoderscited by 90
  5. 2017A Hybrid Deep Learning Architecture for Privacy-Preserving Mobile Analyticscited by 30
  6. 2019Large Margin Softmax Loss for Speaker Verificationcited by 29
  7. 2019Privacy-Preserving Adversarial Representation Learning in ASR: Reality or Illusion?cited by 22
  8. 2019What does a network layer hear? Analyzing hidden representations of end-to-end ASR through speech synthesiscited by 13
  9. 2018To Reverse the Gradient or Not: An Empirical Comparison of Adversarial and Multi-task Learning in Speech Recognitioncited by 10
  10. 2021Configurable Privacy-Preserving Automatic Speech Recognitioncited by 7
  11. 2020Pkwrap: a PyTorch Package for LF-MMI Training of Acoustic Modelscited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2022Are disentangled representations all you need to build speaker anonymization systems?
  2. 2024Exploratory Evaluation of Speech Content Masking
  3. 2026Listen to the Features: Voice Anonymization Driven by Content Embedding Matching over Signal Reconstruction
  4. 2025SEF-MK: Speaker-Embedding-Free Voice Anonymization through Multi-k-means Quantization
  5. 2024Exploring VQ-VAE with Prosody Parameters for Speaker Anonymization
  6. 2026Anonymization, Not Elimination: Utility-Preserved Speech Anonymization
  7. 2022An Ensemble Teacher-Student Learning Approach with Poisson Sub-sampling to Differential Privacy Preserving Speech Recognition
  8. 2025VoxGuard: Evaluating User and Attribute Privacy in Speech via Membership Inference Attacks