Publication
Low-Resource Transliteration for Roman-Urdu and Urdu Using Transformer-Based Models
Muhammad Umer Tariq Butt; Stalin Varanasi; Günter Neumann
In: The Eighth Workshop on Technologies for Machine Translation of Low-Resource Languages. Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT-2025), located at NAACL-2025, April 29 - May 4, New Mexico, USA, Pages 144-153, NAACL, 2025.
Abstract
As the Information Retrieval (IR) field increasingly recognizes the importance of inclusivity, addressing the needs of low-resource languages remains a significant challenge. Transliteration between Urdu and its Romanized form, Roman Urdu, remains underexplored despite the widespread use of both scripts in South Asia. Prior work using RNNs on the Roman-Urdu-Parl dataset showed promising results but suffered from poor domain adaptability and limited evaluation.
We propose a transformer-based approach using the m2m100 multilingual translation model, enhanced with masked language modeling (MLM) pretraining and fine-tuning on both Roman-Urdu-Parl and the domain-diverse Dakshina dataset. To address previous evaluation flaws, we introduce rigorous dataset splits and assess performance using BLEU, characterlevel BLEU, and CHRF.
Our model achieves strong transliteration performance, with Char-BLEU scores of 96.37 for Urdu→Roman-Urdu and 97.44 for Roman-Urdu→Urdu. These results outperform both RNN baselines and GPT-4o Mini and demonstrate the effectiveness of multilingual transfer learning for low-resource transliteration tasks.