mlsanityData to 7 Oct 2026

Making More of Little Data: Improving Low-Resource Automatic Speech Recognition Using Data Augmentation

arXiv 2305.10951 ↗First posted 18 May 2023cs.CL eess.AScited by 30

Not on the map. Why?

Abstract (arXiv)

The performance of automatic speech recognition (ASR) systems has advanced substantially in recent years, particularly for languages for which a large amount of transcribed speech is available. Unfortunately, for low-resource languages, such as minority languages, regional languages or dialects, ASR performance generally remains much lower. In this study, we investigate whether data augmentation techniques could help improve low-resource ASR performance, focusing on four typologically diverse minority languages or language variants (West Germanic: Gronings, West-Frisian; Malayo-Polynesian: Besemah, Nasal). For all four languages, we examine the use of self-training, where an ASR system trained with the available human-transcribed data is used to generate transcriptions, which are then combined with the original data to train a new ASR system. For Gronings, for which there was a pre-existing text-to-speech (TTS) system available, we also examined the use of TTS to generate ASR training data from text-only sources. We find that using a self-training approach consistently yields improved performance (a relative WER reduction up to 20.5% compared to using an ASR system trained on 24 minutes of manually transcribed speech). The performance gain from TTS augmentation for Gronings was even stronger (up to 25.5% relative reduction in WER compared to a system based on 24 minutes of manually transcribed speech). In sum, our results show the benefit of using self-training or (if possible) TTS-generated data as an efficient solution to overcome the limitations of data availability for resource-scarce languages in order to improve ASR performance.

Key facts. Cited by 30 papers in the map (data through 7 Oct 2026). Builds on 24 of its 34 references that are in the map. First posted on arXiv 18 May 2023.

Builds on

24 of 34 references.

References · 24

  1. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,853
  2. 2013An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networkscited by 711
  3. 2020FastSpeech 2: Fast and High-Quality End-to-End Text to Speechcited by 710
  4. 2021XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scalecited by 589
  5. 2019Self-Training for End-to-End Speech Recognitioncited by 128
  6. 2020Self-training and Pre-training are Complementary for Speech Recognitioncited by 98
  7. 2018Rapid Adaptation of Neural Machine Translation to New Languagescited by 82
  8. 2020Iterative Pseudo-Labeling for Speech Recognitioncited by 81
  9. 2019Speech Recognition with Augmented Synthesized Speechcited by 60
  10. 2019Generating Synthetic Audio Data for Attention-Based Speech Recognition Systemscited by 46
  11. 2021Pseudo-Labeling for Massively Multilingual Speech Recognitioncited by 23
  12. 2023Adapting Multilingual Speech Representation Model for a New, Underresourced Language through Multilingual Fine-tuning and Continued Pretrainingcited by 18
12 more
  1. 2022Accented Speech Recognition: Benchmarking, Pre-training, and Diverse Datacited by 15
  2. 2020Neural Representations for Modeling Variation in Speechcited by 14
  3. 2021Voice Conversion Can Improve ASR in Very Low-Resource Settingscited by 13
  4. 2021Magic dust for cross-lingual adaptation of monolingual wav2vec-2.0cited by 12
  5. 2022Quantifying Language Variation Acoustically with Few Resourcescited by 12
  6. 2022Continuous Pseudo-Labeling from the Startcited by 11
  7. 2022Wav2Vec-Aug: Improved self-supervised training with limited datacited by 9
  8. 2021Leveraging pre-trained representations to improve access to untranscribed speech from endangered languagescited by 8
  9. 2022When Is TTS Augmentation Through a Pivot Language Useful?cited by 7
  10. 2022Sample-Efficient Unsupervised Domain Adaptation of Speech Recognition Systems A case study for Modern Greekcited by 7
  11. 2023Leveraging supplementary text data to kick-start automatic speech recognition system development with limited transcriptionscited by 4
  12. 2022Automated speech tools for helping communities process restricted-access corpora for language revival effortscited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 1; 2024: 15; 2025: 7; 2026: 7.

30 papers.

Sort:

By area:

  1. 2023Can Whisper perform speech-based in-context learning?cited by 35
  2. 2024Casablanca: Data and Models for Multidialectal Arabic Speech Recognitioncited by 20
  3. 2024Learn and Don't Forget: Adding a New Language to ASR Foundation Modelscited by 16
  4. 2024Task Arithmetic can Mitigate Synthetic-to-Real Gap in Automatic Speech Recognitioncited by 14
  5. 2025SoundMind: RL-Incentivized Logic Reasoning for Audio-Language Modelscited by 14
  6. 2024The Zeno's Paradox of `Low-Resource' Languagescited by 12
  7. 2024Predicting positive transfer for improved low-resource speech recognition using acoustic pseudo-tokenscited by 11
  8. 2024Enhancing Low-Resource ASR through Versatile TTS: Bridging the Data Gapcited by 11
  9. 2025Data Quality Issues in Multilingual Speech Datasets: The Need for Sociolinguistic Awareness and Proactive Language Planningcited by 10
  10. 2024Improving child speech recognition with augmented child-like speechcited by 7
  11. 2024What is "Typological Diversity" in NLP?cited by 3
  12. 2024Pre-training with Synthetic Patterns for Audiocited by 3
18 more
  1. 2026Speaker-Aware Simulation Improves Conversational Speech Recognitioncited by 2
  2. 2024CMULAB: An Open-Source Framework for Training and Deployment of Natural Language Processing Modelscited by 1
  3. 2024Contrastive Learning from Synthetic Audio Doppelg"angerscited by 1
  4. 2024ManWav: The First Manchu ASR Modelcited by 1
  5. 2024Fast Streaming Transducer ASR Prototyping via Knowledge Distillation with Whispercited by 1
  6. 2025Improving Multilingual Speech Models on ML-SUPERB 2.0: Fine-tuning with Data Augmentation and LID-Aware CTCcited by 1
  7. 2025A Deep Learning Automatic Speech Recognition Model for Shona Languagecited by 1
  8. 2026Doing More with Less: Data Augmentation for Sudanese Dialect Automatic Speech Recognitioncited by 1
  9. 2024TI-ASU: Toward Robust Automatic Speech Understanding through Text-to-speech Imputation Against Missing Speech Modalitycited by 0
  10. 2024Exploring the Impact of Data Quantity on ASR in Extremely Low-resource Languagescited by 0
  11. 2025Methods to Increase the Amount of Data for Speech Recognition for Low Resource Languagescited by 0
  12. 2025Weak Supervision Techniques towards Enhanced ASR Models in Industry-level CRM Systemscited by 0
  13. 2025Evaluating Machine Translation Datasets for Low-Web Data Languages: A Gendered Lenscited by 0
  14. 2026Elderly-Contextual Data Augmentation via Speech Synthesis for Elderly ASRcited by 0
  15. 2026Can Large Language Models Reliably Correct Errors in Low-Resource ASR? A Contamination-Aware Case Study on West Frisiancited by 0
  16. 2026Synthetic Speech, Real Signal: Paralinguistic Preservation and Cross-Lingual Augmentation via Voice Cloningcited by 0
  17. 2026Phoneme-guided TTS augmentation for ASR: A unified pipeline and multilingual evaluationcited by 0
  18. 2026T-SANDHI: Tone Sandhi-aware Adaptive Network with Decoupled Hybrid Injection for Low-resource Taiwanese Hokkien Speech Recognitioncited by 0