Open Language Data Initiative: Advancing Low-Resource Machine Translation for Karakalpak

Mukhammadsaid Mamasaidov, Abror Shopulatov

Open Language Data Initiative: Advancing Low-Resource Machine Translation for Karakalpak: 11 upvotes on Hugging Face Daily Papers, #9 of 12 papers on 2024-09-10. Day-by-day upvote history.

This study presents several contributions for the Karakalpak language: a FLORES+ devtest dataset translated to Karakalpak, parallel corpora for Uzbek-Karakalpak, Russian-Karakalpak and English-Karakalpak of 100,000 pairs each and open-sourced fine-tuned neural models for translation across these languages. Our experiments compare different model variants and training approaches, demonstrating improvements over existing baselines. This work, conducted as part of the Open Language Data Initiative (OLDI) shared task, aims to advance machine translation capabilities for Karakalpak and contribute to expanding linguistic diversity in NLP technologies.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.