mlsanityData to 14 Oct 2026

Reduce, Reuse, Recycle: Is Perturbed Data better than Other Language augmentation for Low Resource Self-Supervised Speech Models

arXiv 2309.12763 ↗First posted 22 Sep 2023eess.AS cs.CL cs.SDcited by 0

Not on the map. Why?

Abstract (arXiv)

Self-supervised representation learning (SSRL) has demonstrated superior performance than supervised models for tasks including phoneme recognition. Training SSRL models poses a challenge for low-resource languages where sufficient pre-training data may not be available. A common approach is cross-lingual pre-training. Instead, we propose to use audio augmentation techniques, namely: pitch variation, noise addition, accented target language and other language speech to pre-train SSRL models in a low resource condition and evaluate phoneme recognition. Our comparisons found that a combined synthetic augmentations (noise/pitch) strategy outperformed accent and language knowledge transfer. Furthermore, we examined the scaling factor of augmented data to achieve equivalent performance to model pre-trained with target domain speech. Our findings suggest that for resource-constrained languages, combined augmentations can be a viable option than other augmentations.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 22 of its 33 references that are in the map. First posted on arXiv 22 Sep 2023.

Builds on

22 of 33 references.

References · 22

  1. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  2. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,905
  3. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,342
  4. 2019SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognitioncited by 1,380
  5. 2021SUPERB: Speech processing Universal PERformance Benchmarkcited by 666
  6. 2021XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scalecited by 601
  7. 2015MUSAN: A Music, Speech, and Noise Corpuscited by 425
  8. 2017The Effectiveness of Data Augmentation in Image Classification using Deep Learningcited by 370
  9. 2019Mockingjay: Unsupervised Speech Representation Learning with Deep Bidirectional Transformer Encoderscited by 198
  10. 2019An Unsupervised Autoregressive Model for Speech Representation Learningcited by 186
  11. 2021Unsupervised Speech Recognitioncited by 148
  12. 2021A Survey on Data Augmentation for Text Classificationcited by 111
10 more
  1. 2019Generative Pre-Training for Speech with Autoregressive Predictive Codingcited by 82
  2. 2020Vector-Quantized Autoregressive Predictive Codingcited by 69
  3. 2020Data Augmenting Contrastive Learning of Speech Representations in the Time Domaincited by 63
  4. 2020Improved Speech Representations with Multi-Target Autoregressive Predictive Codingcited by 23
  5. 2022A Noise-Robust Self-supervised Pre-training Model Based Speech Representation Learning for Automatic Speech Recognitioncited by 22
  6. 2022SPIRAL: Self-supervised Perturbation-Invariant Representation Learning for Speech Pre-Trainingcited by 19
  7. 2021Wav2vec-Switch: Contrastive Learning from Original-noisy Speech Pairs for Robust Speech Recognitioncited by 16
  8. 2021Improving Noise Robustness of Contrastive Speech Representation Learning with Speech Reconstructioncited by 13
  9. 2022SQuId: Measuring Speech Naturalness in Many Languagescited by 10
  10. 2022Wav2Vec-Aug: Improved self-supervised training with limited datacited by 9

Built on by

Not cited yet.