Machine Learning for Endangered Language Documentation (MELD) Lab

The Machine Learning for Endangered Language Documentation (MELD) Lab is directed by Sarah Moeller (smoeller@ufl.edu). It is dedicated to the advancement of natural human language technology, the documentation and description of languages from a computational perspective, and a richer and more accurate representation of all the world’s 7000+ languages in all fields related to artificial intelligence (AI). Its researchers use linguistic insight to improve natural language processing (NLP) and apply NLP to advance linguistic sciences. As our name suggests, much of our work relates in some way to endangered or low-resource languages or genres, but we do much more. We “MELD” together computer science and linguistics! We investigate any question that asks how these two fields can benefit each other. We welcome theoretical work that encompasses refinement through rigorous empirical testing. The potential impact of what we do constantly surprises even us! The research we undertake, the techniques we develop, and the tools we use can be applied to any field that uses data in the form of text or speech, particularly if that data is limited.