Ju Sun and Rui Zhang Receive Funding From Cisco for Federating Learning Research

Ju Sun, PhD, and Rui Zhang, PhD, received funding from the technology corporation Cisco Systems, Inc. to develop federated learning for algorithms to analyze and classify clinical text data. They are also developing new methods to perform federated learning with imbalanced data.

Federated learning is a decentralized approach to train artificial intelligence models. Instead of giving one model all of its training data at once, separate copies of the model are trained in independent sessions using different, smaller sets of data. The trained copies of the model are uploaded to a server and integrated into one, centralized model.

Drs. Sun and  Zhang will develop and validate federated learning for natural language processing (NLP) algorithms, or algorithms that analyze and process human language, to classify clinical text data. The federated learning will then be developed further with a partner at Mayo Clinic where the models will extract phenotype data from cancer patients.

Their research team will also develop new federated learning methods for training models with imbalanced data. Imbalanced data sets have much more training data for some categories than others, leading to misleading results.

This project will build on Drs. Sun and Zhang’s previous work where they developed federated learning for medical images.

Dr. Sun is a full faculty in the Program for Clinical AI and Healthcare Computer Vision lab. Learn more about the lab. Dr. Zhang is co-Director of the Innovative Methods & Data Science (IMDS) program which helps researchers develop new methods for analyzing complex health data. Learn more about the IMDS program.