Network Averaging: Theory and Applications

This lecture by the instructor on May 22, 2020, delves into the concept of averaging networks, exploring both labeled and unlabeled networks. The presentation covers the behavior of averages, statistical methods, and the challenges of network analysis. It discusses the development of 'Statistics 101' for network data objects, the geometry of network spaces, and the computation of Fréchet means. The lecture also touches on practical considerations, such as covariance estimation and large-sample testing theory. Through examples from the 1000 Functional Connectomes Project, the instructor illustrates the application of network averaging in MRI data analysis. The talk concludes by emphasizing the importance of expanding the statistical toolbox to effectively analyze networks.