Jie Xu, Ph.D.

My research interests are machine learning, health informatics, and the intersection of both, with a particular focus on metric learning, federated learning, and privacy-preserving techniques.

I have been working on developing novel computational algorithms for analyzing various kinds of healthcare data, including Electronic Health Records (EHRs), medical and pharmacy claims data, medical imaging data, etc. One specific research direction that I am pursuing is metric learning. It aims to automatically learn a task-specific distance function to effectively calculate the similarity between the input data. One key aspect I have been working on is evaluating the clinical similarity between pairwise patients according to their historical EHR in a federated environment. Besides this, I also work on privacy-preserving technology like differential privacy in order to further protect patients’ information. Another major research topic that I am working on is to identify the potential subtypes of Alzheimer’s disease and use machine learning models to predict the future incidence of AD using administrative EHR in individuals.

0000-0001-5291-5198

• Machine learning and applications