Applied and Interdisciplinary Mathematics Seminar

University of Michigan

Winter 2010
Thursday, April 15, 3:10-4:00pm, 4088 East Hall

Multi-Manifold Data Modeling: Foundations and Applications.

Gilad Lerman

University of Minnesota


Abstract

We present several methods for multi-manifold data modeling, i.e., modeling data by mixtures of possibly intersection manifolds. We focus on algorithms for the special case of hybrid linear modeling, that is, where the underlying manifolds are affine or linear subspaces. We emphasize various theoretical results supporting the performance of some of these algorithms, in particular their robustness to noise and outliers. We demonstrate how such theoretical insights guide us in practical choices. We also present various applications of such algorithms. This is part of various joint works with E. Arias-Castro, S. Atev, G. Chen, A. Szlam, Y. Wang, T. Whitehouse and T. Zhang