Transformer-Based Low-Resource Language Translation: A Study on Standard Bengali to Sylheti
Fuente:
arXiv
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911225689931776 |
|---|---|
| author | Oni, Mangsura Kabir Prama, Tabia Tanzin |
| author_facet | Oni, Mangsura Kabir Prama, Tabia Tanzin |
| contents | Machine Translation (MT) has advanced from rule-based and statistical methods to neural approaches based on the Transformer architecture. While these methods have achieved impressive results for high-resource languages, low-resource varieties such as Sylheti remain underexplored. In this work, we investigate Bengali-to-Sylheti translation by fine-tuning multilingual Transformer models and comparing them with zero-shot large language models (LLMs). Experimental results demonstrate that fine-tuned models significantly outperform LLMs, with mBART-50 achieving the highest translation adequacy and MarianMT showing the strongest character-level fidelity. These findings highlight the importance of task-specific adaptation for underrepresented languages and contribute to ongoing efforts toward inclusive language technologies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_18898 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Transformer-Based Low-Resource Language Translation: A Study on Standard Bengali to Sylheti Oni, Mangsura Kabir Prama, Tabia Tanzin Computation and Language Computers and Society Machine Translation (MT) has advanced from rule-based and statistical methods to neural approaches based on the Transformer architecture. While these methods have achieved impressive results for high-resource languages, low-resource varieties such as Sylheti remain underexplored. In this work, we investigate Bengali-to-Sylheti translation by fine-tuning multilingual Transformer models and comparing them with zero-shot large language models (LLMs). Experimental results demonstrate that fine-tuned models significantly outperform LLMs, with mBART-50 achieving the highest translation adequacy and MarianMT showing the strongest character-level fidelity. These findings highlight the importance of task-specific adaptation for underrepresented languages and contribute to ongoing efforts toward inclusive language technologies. |
| title | Transformer-Based Low-Resource Language Translation: A Study on Standard Bengali to Sylheti |
| topic | Computation and Language Computers and Society |
| url | https://arxiv.org/abs/2510.18898 |