2026· International Conference on Language Resources and Evaluation· pp. 8567-8576· 0 citations· 29 references
Computer Science
TL;DR
A corpus containing both manually curated and synthetically generated sentences for low-resource Indian languages, such as Maithili is contributed and it is demonstrated that, even with a smaller corpus size, high-quality, task-specific data significantly enhance translation accuracy for low-resource Indian languages, such as Maithili.
This work introduces the first-of-its-kind English-Assamese parallel corpus for the translation of Indian court judgments and performs a comprehensive evaluation of state-of-the-art multilingual models in both zero-shot and fine-tuned settings, comparing their performance against commercial systems.
Telem Joyson Singh, Hemanta Baruah, Sanasam Ranbir Singh et al.· International Conference on...· 0 citations
A systematic review of stemming techniques across Afro-Asiatic, Indo-Aryan, Turkic, and Uralic language families, with explicit acknowledgment of coverage limitation, points out the strengths and weaknesses of current approaches and provides insights into potential areas for future investigation and developments in lan...
Sanjiban Sekhar Roy, Mohd Anas, Saravanakumar Kandasamy· Frontiers in Artificial Inte...· 0 citations
MameLoshnLM, the first open-source 8B-parameter language model built specifically for Yiddish, is presented, providing both a foundation for Yiddish NLP and a practical template for language model development in historically rich but digitally underrepresented languages.
Uri Katz, Omer Goldman, Tomasz Limisiewicz et al.· 0 citations
A transferable curation framework for endangered languages, the first sizable Kalabari-English parallel corpus, and baseline experiments that reveal both the promise and the hallucination pitfalls of training on highly constrained, domain-specific data are contributed.
O. T. Olise· WORLD JOURNAL OF INNOVATION...· 0 citations
The combination of domain adaptation and architectural rigour establishes MythoBERT as a foundational model for computational mythology and opens new avenues in cultural heritage NLP research.
SraVaani-1.0 achieves the lowest word error rate (WER) on a large number of language-dataset pairs while remaining competitive with the best-performing systems on high resource while being assessed exclusively on the VAANI benchmark.
Sujith Pulikodan, A. Basu, J. PavanKumar et al.· 1 citation· ⚡1
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
