Democratizing access to natural language processing (NLP) technology is crucial, especially for underrepresented and extremely low-resource languages. Previous research has focused on developing labeled and unlabeled corpora for these languages through online scraping and document translation. While these methods have proven effective and cost-efficient, we have identified limitations in the resulting corpora, including a lack of lexical diversity and cultural relevance to local communities. To address this gap, we conduct a case study on Indonesian local languages. We compare the effectiveness of online scraping, human translation, and paragraph writing by native speakers in constructing datasets. Our findings demonstrate that datasets generated through paragraph writing by native speakers exhibit superior quality in terms of lexical diversity and cultural content. In addition, we present the \datasetname{} benchmark, encompassing 12 underrepresented and extremely low-resource languages spoken by millions of individuals in Indonesia. Our empirical experiment results using existing multilingual large language models conclude the need to extend these models to more underrepresented languages. We release the NusaWrites dataset at https://github.com/IndoNLP/nusa-writes.
@article{arxiv.2309.10661,
title = {NusaWrites: Constructing High-Quality Corpora for Underrepresented and Extremely Low-Resource Languages},
author = {Samuel Cahyawijaya and Holy Lovenia and Fajri Koto and Dea Adhista and Emmanuel Dave and Sarah Oktavianti and Salsabil Maulana Akbar and Jhonson Lee and Nuur Shadieq and Tjeng Wawan Cenggoro and Hanung Wahyuning Linuwih and Bryan Wilie and Galih Pradipta Muridan and Genta Indra Winata and David Moeljadi and Alham Fikri Aji and Ayu Purwarianti and Pascale Fung},
journal= {arXiv preprint arXiv:2309.10661},
year = {2023}
}