SAGE: Spliced-Audio Generated Data for Enhancing Foundational Models in Low-Resource Arabic-English Code-Switched Speech Recognition
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909662483316736 |
|---|---|
| author | Farooq, Muhammad Umar Saz, Oscar |
| author_facet | Farooq, Muhammad Umar Saz, Oscar |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_22143 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SAGE: Spliced-Audio Generated Data for Enhancing Foundational Models in Low-Resource Arabic-English Code-Switched Speech Recognition Farooq, Muhammad Umar Saz, Oscar Computation and Language Sound Audio and Speech Processing 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. |
| title | SAGE: Spliced-Audio Generated Data for Enhancing Foundational Models in Low-Resource Arabic-English Code-Switched Speech Recognition |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2506.22143 |