What are knowledge graphs and how can they support under-resourced languages?

While Knowledge Graphs (KGs) have revolutionized applications for high-resource languages like English and Spanish, the majority of the world’s 7,000 languages remain digitally marginalized. This paper addresses this disparity by presenting a robust framework for building KGs specifically designed for under-resourced languages. Utilizing a task-based methodology, we detail a comprehensive lifecycle covering the creation, hosting, curation, and deployment phases. Our results demonstrate that this structured approach not only enhances the digital presence of minority languages but also provides a scalable architecture for representing multilingual data, text, and media. We conclude with actionable best practices to guide future development in indigenous and under-resourced language technologies.

arXiv