---
title: "Doing More with Less — Mansour 2026"
description: "The paper sets a first Whisper benchmark for this dialect and tests whether cheap data… 57.1% WER on OOOK-Eval. 18 references, cited by 1 paper here."
authors: ["Ayman Mansour"]
year: 2026
venue: "arXiv:2601.06802"
source: "https://arxiv.org/abs/2601.06802"
doi: "10.48550/arxiv.2601.06802"
arxiv: "2601.06802"
licence: "CC BY 4.0 (arXiv)"
licence_url: "http://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/mansour-2026-doing-more-less-data-augmentation/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/mansour-2026-doing-more-less-data-augmentation/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/2/) / Doing More with Less: Data Augmentation for Sudanese Dialect Automatic Speech Recognition

# Doing More with Less: Data Augmentation for Sudanese Dialect Automatic Speech Recognition

**Authors:** Ayman Mansour · **Year:** 2026 · **Venue:** arXiv:2601.06802 · **Licence:** [CC BY 4.0 (arXiv)](http://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2601.06802) · [PDF](https://arxiv.org/pdf/2601.06802) · [DOI 10.48550/arxiv.2601.06802](https://doi.org/10.48550/arxiv.2601.06802) · [arXiv 2601.06802](https://arxiv.org/abs/2601.06802) · [OpenAlex](https://openalex.org/W7123740718)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · [1](https://md.research.elamur.ai/free/en/papers/mansour-2026-doing-more-less-data-augmentation/cited-by/) in our library, of the 140,884 papers whose bibliographies we read

## Summary
Sudanese Arabic has very little transcribed speech, and general-purpose recognizers handle it poorly. The paper sets a first Whisper benchmark for this dialect and tests whether cheap data augmentation can stand in for scarce labels. Whisper Small, Medium and Large-V2 were fine-tuned on about 4 hours of Sudanese speech mixed with an MSA studio corpus. Two ways of adding data were then tried on Small and Medium:
- a teacher model labelled unlabelled broadcast clips, and only labels above a confidence threshold were kept;
- a new Lisan-Sudanese TTS set was built by synthesizing heavily diacritized dialect sentences with the Klaam system.

The strongest system was Whisper-Medium trained on 28.4 hours of real, pseudo-labelled and synthetic audio. It reached 57.1% WER on OOOK-Eval and 51.6% on an out-of-domain holdout. For comparison, zero-shot Whisper scored 78.8% at best, and Arabic models tuned on MSA scored 73.8–123%.

All training ran on free-tier cloud GPUs, and the models and pipelines are public. This makes the work a practical recipe for other dialects with few resources.

## Key points
- **Data:**
  - Sudanese Dialect Speech Dataset: 3,544 clips, about 4 h, 31 speakers.
  - Arabic Speech Corpus: MSA studio recordings.
  - OOOK-Eval: about 2 h of broadcast speech drawn from ADI17.
  - Unlabelled OOOK clips: 17,973 were pseudo-labelled.
  - Lisan-Sudanese TTS: 1,878 sentences, 4.6 h of synthetic audio.
- **Plain fine-tuning (7.25 h of MSA+SDN):** WER on OOOK-Eval was 67.7 (Small), 64.1 (Medium) and 62.8 (Large-V2). On the holdout, Medium was best at 57.5, ahead of Large-V2. The larger model gains little when data is this limited.
- **Pseudo-label filtering:** confidence thresholds of 0.7 and 0.9 were compared. The teacher was trained on a cleaned gold subset only.
- **MSA-tuned models:** the ARBML models looked fairly good on OOOK-Eval (73.8–75.5 WER for Medium/Large) but collapsed on the holdout (140–146 WER). The Common Voice model was the worst overall (123 WER).
- **Compute:** Kaggle P100 and Lightning.ai L40S free tiers, 5,000 steps, LR 1e-5. The best result is about a 45% relative gain over zero-shot on out-of-domain data.

## Interesting ideas
- Confidence-gated pseudo-labels at two cut-offs, compared side by side, show that the threshold is a lever trading label accuracy against coverage and generalization.
- The teacher is trained only on a cleaned gold subset before it labels anything, so its mistakes do not compound through the student.
- Sentences for synthesis are chosen by an annotation-density score (the share of diacritics), keeping only the most fully marked text.
- A second holdout from a different source exposed models whose in-domain scores hid poor generalization.
- Real-but-noisy data (pseudo-labels) and clean-but-artificial data (TTS) helped in complementary ways, and the two worked best combined.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): Sweep the agreement threshold for silver parses from the witness parsers and report the effect on both in-domain and held-out scores, rather than fixing one cut-off.
- **I2** (learn): Train the LLM-teacher-tier models that produce silver data on curated gold treebank data only, and keep silver out of their training loop.
- **I3** (tokens): Rank candidate training sentences by how densely their morphology and lemmas are annotated, and prefer the richest ones when training the tagger and lemmatizer.
- **I4** (answer): Test each reader on an extra set of fables or school texts from an unseen source, so that overfitting to one test style shows up as a control failure.
- **I5** (tree): Mix silver parses of real text with sentences generated from known meaning graphs, whose trees are exact by construction, when training gparse.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

Fields: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/)

Topics: [🎙️ Speech & spoken language](https://md.research.elamur.ai/free/en/topics/speech/) · [🐪 Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/)

## Cite

Cite as: Ayman Mansour. 2026. Doing More with Less: Data Augmentation for Sudanese Dialect Automatic Speech Recognition. arXiv:2601.06802.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{mansour2026doing,
  title = {{Doing More with Less: Data Augmentation for Sudanese Dialect Automatic Speech Recognition}},
  author = {Ayman Mansour},
  year = {2026},
  doi = {10.48550/arxiv.2601.06802},
  eprint = {2601.06802},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2601.06802}
}
```

- [← previous Does Machine "know" interpersonal pragmatics? Evidence from MARBERT's learning of emoji pragmatics in Arabic digital discourse](https://md.research.elamur.ai/free/en/papers/shormani-2026-machine-know-interpersonal-pragmatics-evidence/)
- [next → EPOP: A Benchmark Corpus for Assessing NLP Models on Structured Information Extraction in Plant Health](https://md.research.elamur.ai/free/en/papers/nedellec-2026-epop-benchmark-corpus-assessing-nlp/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/2/)

## References

18 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. A. Conneau, A. Baevski, R. Collobert, A. Mohamed, and M. Auli, “Unsupervised cross-lingual representation learning for speech recognition,” arXiv preprint arXiv:2006.13979, 2020. [↗](https://arxiv.org/abs/2006.13979)
2. A. Radford, J. W. Kim, T. Xu, G. Brockman, C. McLeavey, and I. Sutskever, “Robust speech recognition via large-scale weak supervision,” in International conference on machine learning. PMLR, 2023, pp. 28 492–28 518. [our card](https://md.research.elamur.ai/free/en/papers/radford-2022-robust-speech-recognition-large-scale/)
3. Q. Xu, A. Baevski, T. Likhomanenko, P. Tomasello, A. Conneau, R. Collobert, G. Synnaeve, and M. Auli, “Self-training and pre-training are complementary for speech recognition,” in ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2021, pp. 3030–3034. [find](https://scholar.google.com/scholar?q=Q%2E+Xu%2C+A%2E+Baevski%2C+T%2E+Likhomanenko%2C+P%2E+Tomasello%2C+A%2E+Conneau%2C+R%2E+Collobert%2C+G%2E+Synnaeve%2C+and+M%2E+Auli%2C+%E2%80%9CSelf%2Dtraining+and+pre%2Dtraining+are+complementary+for+speech+recognition%2C%E2%80%9D+in+ICASSP+2021%2D2021+IEE)
4. A. Mansour and W. F. Mukhtar, “End-to-end automatic speech recognition model for the sudanese dialect,” arXiv preprint arXiv:2212.10826, 2022. [↗](https://arxiv.org/abs/2212.10826)
5. J. Kahn, A. Lee, and A. Hannun, “Self-training for end-to-end speech recognition,” in ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2020, pp. 7084–7088. [find](https://scholar.google.com/scholar?q=J%2E+Kahn%2C+A%2E+Lee%2C+and+A%2E+Hannun%2C+%E2%80%9CSelf%2Dtraining+for+end%2Dto%2Dend+speech+recognition%2C%E2%80%9D+in+ICASSP+2020%2D2020+IEEE+International+Conference+on+Acoustics%2C+Speech+and+Signal+Processing+%28ICASSP%29%2E+IEEE%2C+2020%2C+pp)
6. A. Rosenberg, Y. Zhang, B. Ramabhadran, Y. Jia, P. Moreno, Y. Wu, and Z. Wu, “Speech recognition with augmented synthesized speech,” in 2019 IEEE automatic speech recognition and understanding workshop (ASRU). IEEE, 2019, pp. 996–1002. [find](https://scholar.google.com/scholar?q=A%2E+Rosenberg%2C+Y%2E+Zhang%2C+B%2E+Ramabhadran%2C+Y%2E+Jia%2C+P%2E+Moreno%2C+Y%2E+Wu%2C+and+Z%2E+Wu%2C+%E2%80%9CSpeech+recognition+with+augmented+synthesized+speech%2C%E2%80%9D+in+2019+IEEE+automatic+speech+recognition+and+understanding+worksho)
7. M. Bartelds, N. San, B. McDonnell, D. Jurafsky, and M. Wieling, “Making more of little data: Improving low-resource automatic speech recognition using data augmentation,” arXiv preprint arXiv:2305.10951, 2023. [↗](https://arxiv.org/abs/2305.10951)
8. B. Talafha, A. Waheed, and M. Abdul-Mageed, “N-shot benchmarking of whisper on diverse arabic speech recognition,” arXiv preprint arXiv:2306.02902, 2023. [↗](https://arxiv.org/abs/2306.02902) [our card](https://md.research.elamur.ai/free/en/papers/talafha-2023-n-shot-benchmarking-whisper-diverse/)
9. Ö. T. Özyilmaz, M. Coler, and M. Valdenegro-Toro, “Overcoming data scarcity in multi-dialectal arabic asr via whisper fine-tuning,” arXiv preprint arXiv:2506.02627, 2025. [↗](https://arxiv.org/abs/2506.02627) [our card](https://md.research.elamur.ai/free/en/papers/ozyilmaz-2025-overcoming-data-scarcity-multi-dialectal/)
10. N. Halabi and M. Wald, “Phonetic inventory for an arabic speech corpus,” 2016. [find](https://scholar.google.com/scholar?q=N%2E+Halabi+and+M%2E+Wald%2C+%E2%80%9CPhonetic+inventory+for+an+arabic+speech+corpus%2C%E2%80%9D+2016%2E)
11. M. Jarrar, F. A. Zaraket, T. Hammouda, D. M. Alavi, and M. Wählisch, “Lisan: Yemeni, iraqi, libyan, and sudanese arabic dialect corpora with morphological annotations,” in 2023 20th ACS/IEEE International Conference on Computer Systems and Applications (AICCSA). IEEE, 2023, pp. 1–7. [find](https://scholar.google.com/scholar?q=M%2E+Jarrar%2C+F%2E+A%2E+Zaraket%2C+T%2E+Hammouda%2C+D%2E+M%2E+Alavi%2C+and+M%2E+Wa%CC%88hlisch%2C+%E2%80%9CLisan%3A+Yemeni%2C+iraqi%2C+libyan%2C+and+sudanese+arabic+dialect+corpora+with+morphological+annotations%2C%E2%80%9D+in+2023+20th+ACS%2FIEEE+Internat)
12. Y. Ren, C. Hu, X. Tan, T. Qin, S. Zhao, Z. Zhao, and T.-Y. Liu, “Fastspeech 2: Fast and high-quality end-to-end text to speech,” arXiv preprint arXiv:2006.04558, 2020. [↗](https://arxiv.org/abs/2006.04558)
13. S. Shon, A. Ali, Y. Samih, H. Mubarak, and J. Glass, “Adi17: A fine-grained arabic dialect identification dataset,” in ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2020, pp. 8244–8248. [our card](https://md.research.elamur.ai/free/en/papers/salameh-2018-fine-grained-arabic-dialect-identification/)
14. R. Ardila, M. Branson, K. Davis, M. Kohler, J. Meyer, M. Henretty, R. Morais, L. Saunders, F. Tyers, and G. Weber, “Common voice: A massively-multilingual speech corpus,” in Proceedings of the twelfth language resources and evaluation conference, 2020, pp. 4218–4222. [our card](https://md.research.elamur.ai/free/en/papers/ardila-2019-common-voice-massively-multilingual-speech/)
15. A. Ali, P. Bell, J. Glass, Y. Messaoui, H. Mubarak, S. Renals, and Y. Zhang, “The mgb-2 challenge: Arabic multi-dialect broadcast media recognition,” in 2016 IEEE Spoken Language Technology Workshop (SLT). IEEE, 2016, pp. 279–284. [our card](https://md.research.elamur.ai/free/en/papers/ali-2016-mgb-2-challenge-arabic-multi/)
16. T. Wolf, L. Debut, V. Sanh, J. Chaumond, C. Delangue, A. Moi, P. Cistac, T. Rault, R. Louf, M. Funtowicz et al., “Transformers: State-of-the-art natural language processing,” in Proceedings of the 2020 conference on empirical methods in natural language processing: system demonstrations, 2020, pp. 38–45. [our card](https://md.research.elamur.ai/free/en/papers/wolf-2020-transformers-state-art-natural-language/)
17. N. Y. Habash, Introduction to Arabic natural language processing. Morgan & Claypool Publishers, 2010. [find](https://scholar.google.com/scholar?q=N%2E+Y%2E+Habash%2C+Introduction+to+Arabic+natural+language+processing%2E+Morgan+%26+Claypool+Publishers%2C+2010%2E)
18. E. J. Hu, Y. Shen, P. Wallis, Z. Allen-Zhu, Y. Li, S. Wang, L. Wang, W. Chen et al., “Lora: Low-rank adaptation of large language models.” ICLR, vol. 1, no. 2, p. 3, 2022. [our card](https://md.research.elamur.ai/free/en/papers/hu-2021-lora-low-rank-adaptation-large/)