Grambank s typological advances support computational research on diverse languages

Abstract

In spite of increasing attention on lessresourced languages in Natural Language Processing (NLP), equitable access to language technologies and inclusion of diverse languages in the development of these technologies remains a problem (Joshi et al., 2020). This disparity in resources and research attention is pronounced only a handful of the world s approximately 7,000 languages receive the majority of scholarly attention (Blasi et al., 2022). Extending the reach of language technologies to diverse, less-resourced languages is important for tackling the challenges of digital equity and inclusion, and incorporating typological information into language transfer and multilingual learning is an important strategy for doing this. Here we introduce the Grambank typological database as a resource to support efforts that leverage typological features to enhance multilingual NLP.