mlsanityData to 14 Oct 2026

SAGE: Spliced-Audio Generated Data for Enhancing Foundational Models in Low-Resource Arabic-English Code-Switched Speech Recognition

arXiv 2506.22143 ↗First posted 27 Jun 2025cs.CL cs.SD eess.AScited by 2

Not on the map. Why?

Abstract (arXiv)

This paper investigates the performance of various speech SSL models on dialectal Arabic (DA) and Arabic-English code-switched (CS) speech. To address data scarcity, a modified audio-splicing approach is introduced to generate artificial CS speech data. Fine-tuning an already fine-tuned SSL model with the proposed Spliced-Audio Generated (SAGE) data results in an absolute improvement on Word Error Rate (WER) of 7.8% on Arabic and English CS benchmarks. Additionally, an Experience Replay (ER) inspired approach is proposed to enhance generalisation across DA and CS speech while mitigating catastrophic forgetting. Integrating an out-of-domain 3-gram language model reduces the overall mean WER from 31.7% to 26.6%. Few-shot fine-tuning for code-switching benchmarks further improves WER by 4.9%. A WER of 31.1% on Arabic-English CS benchmarks surpasses large-scale multilingual models, including USM and Whisper-large-v2 (both over ten times larger) by an absolute margin of 5.5% and 8.4%, respectively.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 25 of its 39 references that are in the map. First posted on arXiv 27 Jun 2025.

Builds on

25 of 39 references.

References · 25

  1. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,057
  2. 2019Unsupervised Cross-lingual Representation Learning at Scalecited by 3,935
  3. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,905
  4. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,342
  5. 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,384
  6. 2018Experience Replay for Continual Learningcited by 1,095
  7. 2020Unsupervised Cross-lingual Representation Learning for Speech Recognitioncited by 501
  8. 2023Scaling Speech Technology to 1,000+ Languagescited by 411
  9. 2021SpeechBrain: A General-Purpose Speech Toolkitcited by 406
  10. 2023Google USM: Scaling Automatic Speech Recognition Beyond 100 Languagescited by 231
  11. 2022Comparative layer-wise analysis of self-supervised speech modelscited by 158
  12. 2024mHuBERT-147: A Compact Multilingual HuBERT Modelcited by 66
13 more
  1. 2021Multilingual and code-switching ASR challenges for low resource Indian languagescited by 53
  2. 2021QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpuscited by 38
  3. 2020Data Augmentation for End-to-end Code-switching Speech Recognitioncited by 20
  4. 2023N-Shot Benchmarking of Whisper on Diverse Arabic Speech Recognitioncited by 20
  5. 2024Task Arithmetic can Mitigate Synthetic-to-Real Gap in Automatic Speech Recognitioncited by 14
  6. 2023Zero Resource Code-switched Speech Benchmark Using Speech Utterance Pairs For Multiple Spoken Languagescited by 12
  7. 2023Multilingual self-supervised speech representations improve the speech recognition of low-resource African languages with codeswitchingcited by 11
  8. 2023Speech collage: code-switched audio generation by collaging monolingual corporacited by 10
  9. 2024ZAEBUC-Spoken: A Multilingual Multidialectal Arabic-English Speech Corpuscited by 10
  10. 2022Exploring WavLM on Speech Enhancementcited by 9
  11. 2021Wav2vec-S: Semi-Supervised Pre-Training for Low-Resource ASRcited by 7
  12. 2022Optimizing Bilingual Neural Transducer with Synthetic Code-switching Text Generationcited by 5
  13. 2024Progressive unsupervised domain adaptation for ASR using ensemble models and multi-stage trainingcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 1.

2 papers.

  1. 2025Adapting Language Balance in Code-Switching Speechcited by 2
  2. 2026CTC-DID: CTC-Based Arabic dialect identification for streaming applicationscited by 0