Team IndiDataMiner at IndoNLP 2025: Hindi Back Transliteration - Roman to Devanagari using LLaMa

Saurabh Kumar, Dhruvkumar Babubhai Kakadiya, Sanasam Ranbir Singh


Abstract
The increasing use of Romanized typing for Indo-Aryan languages on social media poses challenges due to its lack of standardization and loss of linguistic richness. To address this, we propose a sentence-level back-transliteration approach using the LLaMa 3.1 model for Hindi. Leveraging fine-tuning with the Dakshina dataset, our approach effectively resolves ambiguities in Romanized Hindi text, offering a robust solution for converting it into the native Devanagari script.
Anthology ID:
2025.indonlp-1.15
Volume:
Proceedings of the First Workshop on Natural Language Processing for Indo-Aryan and Dravidian Languages
Month:
January
Year:
2025
Address:
Abu Dhabi
Editors:
Ruvan Weerasinghe, Isuri Anuradha, Deshan Sumanathilaka
Venues:
IndoNLP | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
129–134
Language:
URL:
https://aclanthology.org/2025.indonlp-1.15/
DOI:
Bibkey:
Cite (ACL):
Saurabh Kumar, Dhruvkumar Babubhai Kakadiya, and Sanasam Ranbir Singh. 2025. Team IndiDataMiner at IndoNLP 2025: Hindi Back Transliteration - Roman to Devanagari using LLaMa. In Proceedings of the First Workshop on Natural Language Processing for Indo-Aryan and Dravidian Languages, pages 129–134, Abu Dhabi. Association for Computational Linguistics.
Cite (Informal):
Team IndiDataMiner at IndoNLP 2025: Hindi Back Transliteration - Roman to Devanagari using LLaMa (Kumar et al., IndoNLP 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.indonlp-1.15.pdf