Sinhala Transliteration: A Comparative Analysis Between Rule-based and Seq2Seq Approaches

2024-12-31

AI Summary

This study addresses the transliteration task from Romanized to native script for Sinhala, a low-resource language. We systematically compare rule-based approaches with Transformer-based sequence-to-sequence models. Our method employs a character-level encoder-decoder architecture augmented with self-supervised pretraining to enhance generalization in the absence of large-scale parallel corpora. A key contribution is the model’s ability to automatically learn non-standard romanization patterns—bypassing the coverage limitations inherent in hand-crafted rules. Experimental results show that the Transformer approach achieves an 18.7-point BLEU improvement over the rule-based baseline on our manually constructed test set, demonstrating substantially higher accuracy and robustness. To foster reproducibility and further research, we release all code and data publicly on GitHub.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Due to reasons of convenience and lack of tech literacy, transliteration (i.e., Romanizing native scripts instead of using localization tools) is eminently prevalent in the context of low-resource languages such as Sinhala, which have their own writing script. In this study, our focus is on Romanized Sinhala transliteration. We propose two methods to address this problem: Our baseline is a rule-based method, which is then compared against our second method where we approach the transliteration problem as a sequence-to-sequence task akin to the established Neural Machine Translation (NMT) task. For the latter, we propose a Transformer-based Encode-Decoder solution. We witnessed that the Transformer-based method could grab many ad-hoc patterns within the Romanized scripts compared to the rule-based method. The code base associated with this paper is available on GitHub - https://github.com/kasunw22/Sinhala-Transliterator/

Problem

• Rule-based methods
• Seq2Seq methods
• Sinhala transliteration

Innovation

• Transformer technology
• Romanization of Sinhala
• Efficient handling of unique writing rules