SiPaKosa: A Comprehensive Corpus of Canonical and Classical Buddhist Texts in Sinhala and Pali
This work addresses the problem of limited digital access to historical Buddhist texts for scholars and cultural preservation efforts, though it is incremental as it applies existing methods to new data.
The researchers tackled the lack of comprehensive digital resources for Buddhist texts by creating SiPaKosa, a corpus of 786K sentences and 9.25M words in Sinhala and Pali, and found that proprietary language models outperformed open-source ones by factors of three to six times in perplexity scores ranging from 1.09 to 189.67.
SiPaKosa is a comprehensive corpus of Sinhala and Pali doctrinal texts comprising approximately 786K sentences and 9.25M words, incorporating 16 copyright-cleared historical Buddhist documents alongside the complete web-scraped Tripitaka canonical texts. The corpus was created through high-quality OCR using Google Document AI on historical manuscripts, combined with systematic web scraping of canonical repositories, followed by rigorous quality control and metadata annotation. The corpus is organised into language-specific subcorpora: Sinhala and Mixed Sinhala-Pali. We evaluate the performance of language models using ten pretrained models, with perplexity scores ranging from 1.09 to 189.67 on our corpus. This analysis shows that proprietary models significantly outperform open-source alternatives by factors of three to six times. This corpus supports the pretraining of domain-adapted language models, facilitates historical language analysis, and aids in the development of information retrieval systems for Buddhist scholarship while preserving Sinhala cultural heritage.