Conference on Advances in Data Science: Talk Titles & Abstracts for Invited Speakers

Zhengwu Zhang, University of North Carolina, Chapel Hill

Talk Title: “Tensor Decomposition and its Applications in Brain Network Analysis”

Abstract: Tensor format data is ubiquitous in neuroimaging given its advantages in representing spatial and temporal correlations in the data. In this talk, we are particularly interested in analyzing brain structural connectomes obtained from diffusion MRI using techniques from tensor learning. I will show three interrelated applications: 1) Tensor network PCA for dimension reduction; 2) Population-based analysis of super-high resolution connectomes; And 3) a soft tensor regression. We will show distinguished advantages of these methods in analyzing the brain connectome data over the traditional multivariate ones.