Federated learning for healthcare data analysis: from theory to practice
COM2 Level 4
Executive Classroom, COM2-04-02
Abstract:
There are still many practical barriers to promoting Artificial intelligence (AI) in healthcare applications, such as insufficient training samples, difficulties in data sharing and labeling, etc. To overcome these barriers, federated learning (FL) is a trending framework to enable multi-institutional collaboration in machine learning without sharing raw data. This presentation will discuss our ongoing progress in designing FL algorithms that embrace the unique properties of healthcare data from theory to practice. First, I will present our work on theoretically understanding FL training convergence and generalization using neural tangent kernel, called FL-NTK. Second, I will present our algorithms for tackling two salient issues in multi-institutional healthcare data analysis, i.e., feature heterogeneity and label heterogeneity, motivated by our previous theoretical foundation. Third, I will present our up-to-date progress and promising future directions in FL for healthcare data analysis, including FL for populational disease analysis on graphs and data auditing in FL.
Biodata:
Dr. Xiaoxiao Li is an Assistant Professor at the Department of Electrical and Computer Engineering (ECE) at the University of British Columbia (UBC) starting August 2021. Before joining UBC, Dr. Li was a Postdoc Research Fellow in the Computer Science Department at Princeton University. Dr. Li obtained her PhD degree from Yale University in 2020. Dr. Li's research interests range across the interdisciplinary fields of deep learning and biomedical data analysis, aiming to improve the trustworthiness of AI systems for healthcare. Dr. Li has had over 30 papers published in leading machine learning conferences and journals, including NeurIPS, ICML, ICLR, MICCAI, IPMI, ECCV, IEEE Transactions on Medical Imaging, and Medical Image Analysis. Her work has been recognized with several best paper awards at international conferences. Homepage: https://tea.ece.ubc.ca/.