SEAS Department Website

• Home

• Department Overview
• Undergraduate
• Graduate
• Admissions
• Research
• People
• Careers
• Seminars
  • Financial Engineering Practitioners Seminars
  • IEOR-DRO Seminars
  • The New York Quantitative Finance Seminar
  • The Discrete Math Seminar
  • Financial Engineering Summer Informal Seminar
  • IEOR Seminars
  • Industry Presentations
• News & Events
• Resources

Seminars

Covariance Selection Using Semidefinite Programming

Date: 04-04-2006
Start Time: 1:00pm
End Time: 2:00pm
Speaker: Alex d’Aspremont, Princeton University
Location: Uris 333

Abstract

We address a problem of covariance selection, where we seek a trade-off between high likelihood and the number of non-zero elements in the inverse covariance matrix. After relaxation, the problem is directly amenable to now standard interior-point algorithms for convex optimization, but remains challenging due to its size. We first give some results on the theoretical complexity of this problem, by showing that a recent methodology for non-smooth convex optimization due to Nesterov can be applied to greatly improve the complexity estimate given by interior-point algorithms. We also examine block-coordinate descent algorithms aimed at solving very large, noisy (hence dense) instances.

Joint work with O. Banerjee and L. El Ghaoui at University of California, Berkeley.

Bio

Alexandre d'Aspremont is an assistant professor at the department of operations research and financial engineering at Princeton University. He holds dual Ph.D.s from the École Polytechnique and Stanford University. His research is focused on financial engineering and applications of convex programming to finance, statistics, and engineering.