Preethi Jyothi

简介

Speech and NLP technologies have advanced remarkably in recent years. However, they do not work equally well across users from diverse demographics, particularly in India where users span a wide variety of speech accents, dialects and languages. My research focuses on building speech and language technologies for such low-resource settings. A current key challenge of interest is how to enable effective transfer of knowledge from pretrained LLMs to new domains, and how to adapt large language models in parameter-efficient and cost-effective ways.