Presentation:
“(Quickly) Learning to Optimize Wireless Network Management”
Machine learning has shown enormous promise for revolutionizing many industries, telecommunications included. By collecting and learning from historical data on network usage, equipment failures, and more, telecommunications network operators and equipment vendors can improve their operational efficiency and reliability, ultimately providing a better end-user experience, potentially at a reduced cost. Yet realizing this promise requires much more than blindly applying existing machine learning techniques to networking problems. Utilizing learning-based approaches for network management, for example, requires machine learning algorithms that can rapidly adapt with network changes (sometimes at the millisecond timescale) and that can operate across multiple network layers, ranging from the physical layer up to the application layer. Many existing machine learning algorithms generalize poorly to data outside their training datasets; thus, when applied to problems in highly variable networks where user demands can vary significantly over time, they often require large amounts of network data, which may not be available in practice, to successfully adapt to real network settings. I will present some of my research group’s work on designing new machine learning algorithms for communication network management. The key insight of our work is that existing knowledge of network operations can be combined with the learning algorithm to accelerate its ability to manage the network in any given scenario.
Biography:
Carlee Joe-Wong is the Robert E. Doherty Career Development Professor of Electrical and Computer Engineering at Carnegie Mellon University. She received her A.B. degree (magna cum laude) in Mathematics, and M.A. and Ph.D. degrees in Applied and Computational Mathematics, from Princeton University in 2011, 2013, and 2016, respectively. Her research interests lie in optimizing various types of networked systems, in particular applications of machine learning and economics to computing and communication networks. From 2013 to 2014, Carlee was the Director of Advanced Research at DataMi, a startup she co-founded from her research on mobile data pricing. Her work has received best paper and poster awards at several conferences, including IEEE INFOCOM, ACM/IEEE IPSN, ACM SIGMETRICS, and IEEE ICDCS. She received the NSF CAREER award in 2018, the Army Young Investigator award in 2019, and the Department of Energy Early Career Research Program Award in 2024.
