Applications of Sparse Signal Representations: From Signal Processing to Finance

Event Date: February 23, 2011
Speaker: Dmitry Malioutov
Speaker Affiliation: DRW Trading Group, Chicago
Sponsor: Communications, Networking, Signal & Image Processing
Time: 5:00 PM
Location: EE 118
Contact Name: Prof. Ilya Pollak
Contact Phone: (765) 49-45916
Contact Email: ipollak@purdue.edu
Open To: ACCEPTABLE FOR ECE 694A

 

After a brief review of sparse signal representation, we consider its practical applications: first we will look at source localization in sensor arrays and describe a method that produces highly sparse spatial spectra via second order cone programming.  Next we will overview several applications in quantitative finance: construction of sparse Markowitz portfolios, structured risk modeling, and sparsity in statistical arbitrage.  Time permitting we will also touch upon the role of sparsity in designing robust estimators.

 

BIO:

 

Dmitry Malioutov is a researcher in the Algorithmic Trading Group at DRW, Chicago. He obtained his Ph.D. from MIT EECS in 2008, where his research focused on message-passing schemes for inference in graphical models, and sparse signal representation.  Prior to joining DRW Dmitry held a postdoctoral position at Microsoft Research, Cambridge, UK.  His current interests include statistical risk modeling, structured high-dimensional statistical inference, and convex optimization.