Distinguished Lecture Series

Ultra-sparse Matrix Normal Models of Multiway Data

Professor Alfred Hero

John H. Holland Distinguished University Professor of Electrical Engineering and Computer Science, University of Michigan
R. Jamison and Betty Williams Professor of Engineering, University of Michigan

Biography:
Alfred Hero is the John H. Holland Distinguished University Professor of Electrical Engineering and Computer Science and the R. Jamison and Betty Williams Professor of Engineering at the University of Michigan, Ann Arbor. He is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE) and the Society for Industrial and Applied Mathematics (SIAM). He is a recipient of the Fourier Award in Signal Processing from the IEEE. He is a Section Editor of the SIAM Journal on Mathematics of Data Science and a Senior Editor of the IEEE Journal on Selected Topics in Signal Processing.

(Poster) (Photos)

Date:

3 February 2021 (Wednesday) (Original scheduled on 16 December 2020)

Time:

10:00-11:00 a.m. GMT+8 (Hong Kong Time)

Venue:

Online via Zoom (Meeting ID: 935 8469 3865)

Abstract

Modeling multi-way data is important for applications involving multi-indexed observables, e.g., hyperpsectral data that is indexed over spatial, frequency, and temporal dimensions. The sparse matrix normal model is a multivariate Gaussian representation that expresses the covariance matrix as a Kronecker product of sparse lower dimensional covariances. This model is equivalent to assuming the conditional dependencies of the covariates can be represented as a direct-product graph with few edges. We will present an alternative framework based on Cartesian product graph representation and Kronecker sums that leads to ultra-sparse and generative models for multi-way data.

Sponsored by: Joint Research Institute for Applied Mathematics

Supported by:

All are welcome