Faculty Candidate Talk: Ehsan Elhamifar
Sparse Modeling for High-Dimensional Multi-Manifold Data Analysis
One of the most fundamental challenges facing scientists and engineers across different fields, such as computer vision, robotics, bioinformatics and speech/image processing, is the large amounts of high-dimensional data that need to be analyzed and understood. In this talk, I present provably correct and efficient algorithms, based on the sparse representation theory, for the analysis of high-dimensional datasets by exploiting their underlying low-dimensional structures. I talk about algorithms for the two fundamental problems of clustering and subset selection in unions of subspaces and discuss the robustness of the algorithms to data nuisances. I show that these tools effectively advance the state-of-the-art data analysis in a wide range of important real-world problems, such as segmentation of motions in videos, clustering of images of objects, energy disaggregation and learning nonlinear dynamical models.
Ehsan Elhamifar is a postdoctoral scholar in the department of Electrical Engineering and Computer Sciences at the University of California, Berkeley. He obtained his PhD in Electrical and Computer Engineering from the Johns Hopkins University. Ehsan is broadly interested in developing provably correct and efficient data analysis algorithms that can address challenges of complex and large-scale high-dimensional datasets. Specifically, he focuses on the intrinsic low-dimensionality of real data and uses tools from convex geometry and analysis, sparse / low-rank representation, high-dimensional statistics and graph theory to develop such algorithms. Ehsan obtained MS and MSE degrees in Electrical Engineering and Applied Mathematics and Statistics, respectively, from Sharif University of Technology in Iran and the Johns Hopkins University.