[Colloquium] Summer Statistics Seminar: ANDREAS ARGYRIOU, Monday, June 27, 2011, at 4:00 PM, in Eckhart 110

[Lek-Heng Lim]

The University of Chicago
Department of Statistics
Seminar Series

ANDREAS ARGYRIOU
Toyota Technological Institute at Chicago
Structured Sparsity, Proximal Optimization and Multi-Task Learning

MONDAY, June 27, 2011, at 4:00 PM
110 Eckhart Hall, 5734 S. University Avenue

ABSTRACT

In recent years, there has been increasing interest in problems
exhibiting structured sparsity, and methods such as Lasso and Group Lasso
have been used extensively to this end. I will discuss a more general
framework for structured sparsity which can replace ad hoc approaches. It
provides a straightforward way of favoring prescribed sparsity patterns,
such as orderings, contiguous regions and overlapping groups, among
others. I will also present a first-order optimization method for solving
nonsmooth convex programs like these. This method builds on fixed point
and successive approximations theory. Finally, I will discuss an approach
for multitask learning and matrix completion and a related algorithm.


OCCASIONAL SUMMER STATISTICS SEMINAR
Mondays at 4:00 PM, in Eckhart 110

Jun 27: Andreas Argyriou (TTI)
Structured Sparsity, Proximal Optimization and Multi-Task Learning

Jul 11: Thomas Schultz (Chicago)
Symmetric Tensor Fields in Diffusion MRI and Computer Vision

Jul 25: Yuan Yao (Peking)
Geometric and Topological Methods for Data Science

Aug 08: Sonja Petrovic (Penn State)
Algebraic Statistics for Random Graph Models

Aug 15: Suvrit Sra (MPI Tubingen)
Metric Nearness: The Problem, and Recent Advances

Sep 12: Shmuel Friedland (UIC)
Completion of Missing Entries in Matrices and Tensors



For further information, please contact Lek-Heng Lim at 773.702.4263 or 
send email (lekheng at galton.uchicago.edu). If you wish to subscribe to our 
email list, please visit the following website: 
https//lists.uchicago.edu/web/info/statseminars.