Abstract: Recent advances in artificial intelligence (AI) have led to disruptive innovations such as smart manufacturing, mobile healthcare, and autonomous driving. However, training robust and accurate AI models requires large and diverse datasets, which are usually generated at the edge, outside a centralized cloud. Moving the distributed data to a centralized cloud raises both privacy and communication efficiency concerns. This motivates federated learning, a promising learning paradigm for collaborative AI without centralizing data, but there are still several unsolved challenges.

 

In this talk, I will discuss our work on trustworthy and distributed data analytics for intelligent edge. I would cover several core challenges of federated learning, including lack of rigorous privacy guarantee and expensive communication over wireless links. I will also present several of our work on addressing these challenges and providing rigorous differential privacy guarantee and high model accuracy in practice. I will end with some other projects on AI, security and privacy, and networked systems in my group.

Short Bio: Yanmin Gong received her M.S. degree in electrical engineering from Tsinghua University, China, in 2012, and the Ph.D. degree in electrical and computer engineering from the University of Florida, USA, in 2016. She is currently an assistant professor in electrical and computer engineering at UT San Antonio. Her research interests lie at the intersection of machine learning, cybersecurity, and networking systems. She is a recipient of the NSF CAREER Award, the NSF CRII Award, IEEE Computer Society TCSC Early Career Researchers Award for Excellence in Scalable Computing, Rising Star in Networking and Communications Award by IEEE ComSoc N2Women, Distinguished TPC Member Award of IEEE INFOCOM, and Best Paper Award at IEEE GLOBECOM. She is currently an Editor of the IEEE Wireless Communications and IEEE Senior Member. Her research has been supported by NSF, Air Force Research Lab and Army Research Office.