My research develops trustworthy artificial intelligence methods for Alzheimer's disease and related dementia research, spanning generative modeling (for imputing and harmonizing medical imaging data across sites), formal verification of algorithmic fairness, and uncertainty quantification and safety guardrailing for large language models — with the goal of building medical AI systems that are provably reliable, fair, and safe for real clinical and research use.
Fairness auditing and uncertainty quantification, so that models know when they don't know and treat patient groups equitably.
Large language models (LLMs) and multimodal foundation models that reason across text, images, and structured health data, and generative AI whose outputs are reliable, well-calibrated, and evaluable.
Medical imaging, longitudinal health data, and Alzheimer's disease and related dementias (AD/ADRD).