---
title: "Wave2Word@DravidianLangTech 2026: WhisTam: A unified"
description: "As an entry to the DravidianLangTech@ACL 2026 shared task, the authors present WhisTam, one… 17 references, cited by 1 paper here."
authors: ["Ruwad Naswan", "Shadab Tanjeed Ahmad"]
year: 2026
venue: "Proceedings of the Sixth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages"
source: "https://aclanthology.org/2026.dravidianlangtech-1.70/"
doi: "10.18653/v1/2026.dravidianlangtech-1.70"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/naswan-2026-wave2word-dravidianlangtech-2026-whistam-unified/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/naswan-2026-wave2word-dravidianlangtech-2026-whistam-unified/)

[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/19/) / Wave2Word@DravidianLangTech 2026: WhisTam: A unified framework for dialect based Tamil speech recognition and classification

# Wave2Word@DravidianLangTech 2026: WhisTam: A unified framework for dialect based Tamil speech recognition and classification

**Authors:** [Ruwad Naswan](https://md.research.elamur.ai/free/en/authors/ruwad-naswan/) and Shadab Tanjeed Ahmad · **Year:** 2026 · **Venue:** [Proceedings of the Sixth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages](https://md.research.elamur.ai/free/en/venues/dravidianlangtech/) ([Workshop on Speech, Vision, and Language Technologies for Dravidian Languages 2026](https://md.research.elamur.ai/free/en/venues/dravidianlangtech/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.dravidianlangtech-1.70/) · [PDF](https://aclanthology.org/2026.dravidianlangtech-1.70.pdf) · [DOI 10.18653/v1/2026.dravidianlangtech-1.70](https://doi.org/10.18653/v1/2026.dravidianlangtech-1.70) · [OpenAlex](https://openalex.org/W7166862901)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · 1 in our library, of the 149,439 papers whose bibliographies we read

## Summary
Speech recognition for Tamil is hindered by strong regional differences in pronunciation, vocabulary and syntax, and by the small amount of dialect-annotated audio. As an entry to the DravidianLangTech@ACL 2026 shared task, the authors present WhisTam, one system that both transcribes dialectal Tamil and assigns each recording to one of four regions: Northern, Southern, Western or Central. It starts from vasista22/whisper-tamil-medium and adds a Regional Adapter. During training, a learned dialect embedding is projected, passed through LayerNorm and added to every encoder frame. At inference this embedding is removed. A separate head pools the encoder states with attention and predicts the dialect. The transcription and classification losses are optimized together. Training used the 9.22-hour Tamil Dialect Speech Dataset with extensive augmentation. On the hidden test set the system reached macro F1 0.53 and WER 0.55. These scores are far below the validation results, which shows how hard it is to generalize across Tamil dialects. The work also shows that multi-task adaptation can be added to a pretrained model at low cost for low-resource speech.

## Key points
- **Base model and parameters:** WhisTam builds on vasista22/whisper-tamil-medium (Whisper medium) and adds only a small number of parameters. The training loss is L_ASR + α·L_dialect.
- **Data:** The Tamil Dialect Speech Dataset contains 9.22 h of training audio and 2.05 h of test audio at 16 kHz, with both spontaneous and read speech. The training data was split 90/10 into train and validation sets using stratified sampling.
- **Validation results:** WER averaged 37.17% across dialects, ranging from 30.23% for Central to 41.68% for Western. Dialect macro F1 was 0.97.
- **Hidden test results:** WER was 55.0%, ranking 3rd in the ASR subtask. Macro F1 was 0.53, ranking 2nd in the dialect classification subtask.
- **Embeddings and code:** t-SNE plots show clear dialect clusters after training, whereas the pretrained encoder shows none. The code is published at github.com/rwd51/DravidianLangTech-Wave2Word.

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

## Institutions

- [Islamic University of Technology](https://md.research.elamur.ai/free/en/universities/islamic-university-of-technology/) — University, Dhaka, Bangladesh · 10 works · 555 citations
- [Bangladesh University of Engineering and Technology](https://md.research.elamur.ai/free/en/universities/bangladesh-university-of-engineering-and-technology/) — University, Dhaka, Bangladesh · 7 works · 308 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🔊 Signal Processing](https://md.research.elamur.ai/free/en/shelves/computer-science/signal-processing/) · 493 works

🧭 Topics (1): [🎙️ Speech & spoken language](https://md.research.elamur.ai/free/en/topics/speech/) · 519 works

## Cite

Cite as: Ruwad Naswan and Shadab Tanjeed Ahmad. 2026. Wave2Word@DravidianLangTech 2026: WhisTam: A unified framework for dialect based Tamil speech recognition and classification. Proceedings of the Sixth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages.

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

```bibtex
@inproceedings{naswan2026wave2word,
  title = {{Wave2Word@DravidianLangTech 2026: WhisTam: A unified framework for dialect based Tamil speech recognition and classification}},
  author = {Ruwad Naswan and Shadab Tanjeed Ahmad},
  booktitle = {Proceedings of the Sixth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages},
  year = {2026},
  doi = {10.18653/v1/2026.dravidianlangtech-1.70},
  url = {https://doi.org/10.18653/v1/2026.dravidianlangtech-1.70}
}
```

- [← previous VVTE at ATE-IT: From Candidates to Terms: Hybrid Italian ATE with Dependency Heuristics, Gemini, and Random Forest Filtering](https://md.research.elamur.ai/free/en/papers/vandervoort-2026-vvte-ate-it-candidates-terms/)
- [next → Weakly Supervised Named Entity Recognition for Historical Texts](https://md.research.elamur.ai/free/en/papers/sorbi-2026-weakly-supervised-named-entity-recognition/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/19/)

## References

17 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Zrar Khald Abdul and Abdulbasit K. Al-Talabani. 2022. Mel frequency cepstral coefficient and its applications: A review. IEEE Access, 10:122136–122158. [find](https://scholar.google.com/scholar?q=Zrar+Khald+Abdul+and+Abdulbasit+K%2E+Al%2DTalabani%2E+2022%2E+Mel+frequency+cepstral+coefficient+and+its+applications%3A+A+review%2E+IEEE+Access%2C+10%3A122136%E2%80%93122158%2E)
2. Muzaffar Ahmad Dar and Jagalingam Pushparaj. 2025. Machine learning and deep learning approaches for accent recognition: A review. IEEE Access, 13:51527–51550. [find](https://scholar.google.com/scholar?q=Muzaffar+Ahmad+Dar+and+Jagalingam+Pushparaj%2E+2025%2E+Machine+learning+and+deep+learning+approaches+for+accent+recognition%3A+A+review%2E+IEEE+Access%2C+13%3A51527%E2%80%9351550%2E)
3. Alexei Baevski, Yuhao Zhou, Abdelrahman Mohamed, and Michael Auli. 2020. [wav2vec 2.0: A framework for self-supervised learning of speech representations](https://md.research.elamur.ai/free/en/papers/baevski-2020-wav2vec-2-0-framework-self/). In Advances in Neural Information Processing Systems, volume 33, pages 12449–12460. Curran Associates, Inc.
4. Laurent Besacier, Etienne Barnard, Alexey Karpov, and Tanja Schultz. 2014. Automatic speech recognition for under-resourced languages: A survey. Speech Communication, 56:85–100. [find](https://scholar.google.com/scholar?q=Laurent+Besacier%2C+Etienne+Barnard%2C+Alexey+Karpov%2C+and+Tanja+Schultz%2E+2014%2E+Automatic+speech+recognition+for+under%2Dresourced+languages%3A+A+survey%2E+Speech+Communication%2C+56%3A85%E2%80%93100%2E)
5. B. Bharathi, Bharathi Raja Chakravarthi, Shunmuga Priya Muthusamy Chinnan, S. Saranya, and S. Suhasini. 2026. Findings in tamil dialect speech recognition and classification. In Proceedings of the Sixth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages. Association for Co… [find](https://scholar.google.com/scholar?q=B%2E+Bharathi%2C+Bharathi+Raja+Chakravarthi%2C+Shunmuga+Priya+Muthusamy+Chinnan%2C+S%2E+Saranya%2C+and+S%2E+Suhasini%2E+2026%2E+Findings+in+tamil+dialect+speech+recognition+and+classification%2E+In+Proceedings+of+the+Six)
6. B. Bharathi, S. Saranya, P. Vijayalakshmi, and T. Nagarajan. 2025. Multi-dialect speech corpus creation for enhancing tamil automatic speech recognition. Circuits, Systems, and Signal Processing, pages 1– 19. [find](https://scholar.google.com/scholar?q=B%2E+Bharathi%2C+S%2E+Saranya%2C+P%2E+Vijayalakshmi%2C+and+T%2E+Nagarajan%2E+2025%2E+Multi%2Ddialect+speech+corpus+creation+for+enhancing+tamil+automatic+speech+recognition%2E+Circuits%2C+Systems%2C+and+Signal+Processing%2C+page)
7. Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, and Michael Auli. 2020. Unsupervised cross-lingual representation learning for speech recognition. Preprint, arXiv:2006.13979. [find](https://scholar.google.com/scholar?q=Alexis+Conneau%2C+Alexei+Baevski%2C+Ronan+Collobert%2C+Abdelrahman+Mohamed%2C+and+Michael+Auli%2E+2020%2E+Unsupervised+cross%2Dlingual+representation+learning+for+speech+recognition%2E+Preprint%2C+arXiv%3A2006%2E13979%2E)
8. Shashidhar G. Koolagudi, Deepika Rastogi, and K. S. Sekhara Rao. 2012. Identification of language using mel-frequency cepstral coefficients (mfcc). Procedia Engineering, 38:3391–3398. [find](https://scholar.google.com/scholar?q=Shashidhar+G%2E+Koolagudi%2C+Deepika+Rastogi%2C+and+K%2E+S%2E+Sekhara+Rao%2E+2012%2E+Identification+of+language+using+mel%2Dfrequency+cepstral+coefficients+%28mfcc%29%2E+Procedia+Engineering%2C+38%3A3391%E2%80%933398%2E)
9. A. Madhavaraj and Ramakrishnan Angarai Ganesan. 2022. Data and knowledge-driven approaches for multilingual training to improve the performance of speech recognition systems of indian languages. Preprint, arXiv:2201.09494. — [arXiv:2201.09494](https://arxiv.org/abs/2201.09494) [find](https://scholar.google.com/scholar?q=A%2E+Madhavaraj+and+Ramakrishnan+Angarai+Ganesan%2E+2022%2E+Data+and+knowledge%2Ddriven+approaches+for+multilingual+training+to+improve+the+performance+of+speech+recognition+systems+of+indian+languages%2E+Prepr)
10. M. Nanmalar, S. Johanan Joysingh, P. Vijayalakshmi, and T. Nagarajan. 2024a. A feature engineering approach for literary and colloquial tamil speech classification using 1d-cnn. Speech Commun., 173:103254. [find](https://scholar.google.com/scholar?q=M%2E+Nanmalar%2C+S%2E+Johanan+Joysingh%2C+P%2E+Vijayalakshmi%2C+and+T%2E+Nagarajan%2E+2024a%2E+A+feature+engineering+approach+for+literary+and+colloquial+tamil+speech+classification+using+1d%2Dcnn%2E+Speech+Commun%2E%2C+173%3A10)
11. M. Nanmalar, S. Johanan Joysingh, P. Vijayalakshmi, and T. Nagarajan. 2024b. A feature engineering approach for literary and colloquial tamil speech classification using 1d-cnn. Speech Commun., 173:103254. [find](https://scholar.google.com/scholar?q=M%2E+Nanmalar%2C+S%2E+Johanan+Joysingh%2C+P%2E+Vijayalakshmi%2C+and+T%2E+Nagarajan%2E+2024b%2E+A+feature+engineering+approach+for+literary+and+colloquial+tamil+speech+classification+using+1d%2Dcnn%2E+Speech+Commun%2E%2C+173%3A10)
12. M. Nanmalar, P. Vijayalakshmi, and T. Nagarajan. 2019. Literary and colloquial dialect identification for tamil using acoustic features. [find](https://scholar.google.com/scholar?q=M%2E+Nanmalar%2C+P%2E+Vijayalakshmi%2C+and+T%2E+Nagarajan%2E+2019%2E+Literary+and+colloquial+dialect+identification+for+tamil+using+acoustic+features%2E)
13. TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON), pages 1303–1306. [find](https://scholar.google.com/scholar?q=TENCON+2019+%2D+2019+IEEE+Region+10+Conference+%28TENCON%29%2C+pages+1303%E2%80%931306%2E)
14. M. Nanmalar, P. Vijayalakshmi, and T. Nagarajan. 2022. Literary and colloquial tamil dialect identification. Circuits, Systems, and Signal Processing, 41:4004 – 4027. [find](https://scholar.google.com/scholar?q=M%2E+Nanmalar%2C+P%2E+Vijayalakshmi%2C+and+T%2E+Nagarajan%2E+2022%2E+Literary+and+colloquial+tamil+dialect+identification%2E+Circuits%2C+Systems%2C+and+Signal+Processing%2C+41%3A4004+%E2%80%93+4027%2E)
15. Vineel Pratap, Anuroop Sriram, Paden Tomasello, Awni Hannun, Vitaliy Liptchinsky, Gabriel Synnaeve, and Ronan Collobert. 2020. Massively multilingual asr: 50 languages, 1 model, 1 billion parameters. Preprint, arXiv:2007.03001. — [arXiv:2007.03001](https://arxiv.org/abs/2007.03001) [find](https://scholar.google.com/scholar?q=Vineel+Pratap%2C+Anuroop+Sriram%2C+Paden+Tomasello%2C+Awni+Hannun%2C+Vitaliy+Liptchinsky%2C+Gabriel+Synnaeve%2C+and+Ronan+Collobert%2E+2020%2E+Massively+multilingual+asr%3A+50+languages%2C+1+model%2C+1+billion+parameters%2E+)
16. Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey, and Ilya Sutskever. 2022. [Robust speech recognition via large-scale weak supervision](https://md.research.elamur.ai/free/en/papers/radford-2022-robust-speech-recognition-large-scale/). Preprint, arXiv:2212.04356.
17. Hem Das and Utpal Bhattacharjee. 2024. Assamese dialect identification using static and dynamic features from vowel. Journal of Advances in Information Technology, 15:306–321. [find](https://scholar.google.com/scholar?q=Hem+Das+and+Utpal+Bhattacharjee%2E+2024%2E+Assamese+dialect+identification+using+static+and+dynamic+features+from+vowel%2E+Journal+of+Advances+in+Information+Technology%2C+15%3A306%E2%80%93321%2E)

*Updated 31 October 2026*