Low-Resource Language Technology 

Examining the availability and efficacy of large language models (LLMs) for low-resourced languages (LRLs). How can we develop resources and methods that make LLMs more effective for LRL users? How do policy actors and science communities impact the development of tools and resources within specific LRLs?

The rapid success of large language models (LLMs) in high-resourced languages such as English has not equally benefited the thousands of low-resourced languages (LRLs) that have mostly been excluded from consideration in LLM development. We are examining the availability and efficacy of LLMs for LRLs. We want to develop resources and methods that make LLMs more effective for LRL users. In addition, we are exploring how policy actors and science communities impact the development of tools and resources within specific LRLs. For example, we have examined the availability of natural language processing tools and digital corpora and the work of scientific communities in low-resourced Sino-Tibetan languages.