Biostatistics and Medical Informatics Seminar with Didong Li, University of North Carolina-Chapel Hill

Dr. Li will discuss the impact of AI on statistics. He will discuss models from his lab, including theoretical results for generative models, a method to use embeddings from LLMs to enhance high-dimensional hypothesis testing, an extensions to genetic studies that include curating annotations for 8.9B variants for downstream tasks. Finally, he will introduce STimage-14KM, a publicly available transriptomics dataset curated by his group that has facilitated the training foundation data models.