Learning over Graphs

Abstract：

 Network science deals with issues related to the aggregation, processing,and diffusion of information over graphs. While interactions among agents can bestudied from the perspective of cluster formations, degrees of connectivity, andsmall-world effects, it is the possibility of having agents interact dynamically witheach other, and influence each other's behavior, that opens up a plethora of notablepossibilities and challenges. For example, the examination of how local interactionsinfluence global behavior can lead to a broader understanding of how localizedinteractions in the social sciences, life sciences, and system sciences influence theevolution of the respective multi-agent networks.In this presentation, we explain how agents can learn from dispersed informationand solve inference and learning tasks of varying degrees of complexity throughlocalized processing and diffusion learning. The presentation explains howinformation is diffused over graphs, how beliefs are formed, and how the graphtopology influences performance. Examples will be considered in the context ofsocial learning, teamwork, and distributed optimization.

Bio: 

Ali H. Sayed served as Dean of Engineering at EPFL, Switzerland.during 2017-2025, where he also directs the Adaptive Systems Laboratory. He servedbefore as Distinguished Professor and Chair of Electrical Engineering at UCLA.where he was a faculty member during the period 1996-2021. He is a member of theUS National Academy of Engineering (NAE) and The World Academy of Sciences(TWAS). He served as President of the IEEE Signal Processing Society in 2018 and2019. An author of over 660 scholarly publications and 10 books, his researchinvolves several areas including adaptation and learning theories, statisticalinference, and multi-agent systems. His work has been recognized with several majorawards including the 2022 IEEE Fourier Technical Field Award and the 2020 IEEEWiener Society Award.

Tencent Meeting：409-433-182