BEGIN:VCALENDAR
METHOD:PUBLISH
PRODID:-//Apple Computer\, Inc//iCal 1.0//EN
X-WR-CALNAME;VALUE=TEXT:USC
VERSION:2.0
BEGIN:VEVENT
DESCRIPTION:Speaker: Geert Leus, Delft University of Technology
Talk Title: Sparsity-Cognizant Total Least-Squares for Perturbed Compressive Sampling
Abstract: Solving linear regression problems based on the total\n
least-squares (TLS) criterion has well-documented merits in various\n
applications, where perturbations appear both in the data vector as well as\n
in the regression matrix. However, existing TLS approaches do not account\n
for sparsity possibly present in the unknown vector of regression\n
coefficients. On the other hand, sparsity is the key attribute exploited by\n
modern compressive sampling and variable selection approaches to linear\n
regression, which include noise in the data, but do not account for\n
perturbations in the regression matrix. In this presentation, we fill this\n
gap by formulating and solving TLS optimization problems under sparsity\n
constraints. Near-optimum and reduced-complexity suboptimum sparse (S-) TLS\n
algorithms are developed to address the perturbed compressive sampling (and\n
the related dictionary learning) challenge, when there is a mismatch between\n
the true and adopted bases over which the unknown vector is sparse. The\n
novel S-TLS schemes also allow for perturbations in the regression matrix of\n
the least-absolute selection and shrinkage selection operator (Lasso), and\n
endow TLS approaches with ability to cope with sparse, under-determined\n
errors-in-variables models. Interesting generalizations can further exploit\n
prior knowledge on the perturbations to obtain novel weighted and structured\n
S-TLS solvers. Analysis and simulations demonstrate the practical impact of\n
S-TLS in calibrating the mismatch effects of contemporary grid-based\n
approaches to cognitive radio sensing, and robust direction-of-arrival\n
estimation using antenna arrays.\n
Biography: Geert Leus was born in Leuven, Belgium, in 1973. He received the\n
electrical engineering degree and the PhD degree in applied sciences from\n
the Katholieke Universiteit Leuven, Belgium, in June 1996 and May 2000,\n
respectively. He has been a Research Assistant and a Postdoctoral Fellow of\n
the Fund for Scientific Research - Flanders, Belgium, from October 1996 till\n
September 2003. During that period, Geert Leus was affiliated with the\n
Electrical Engineering Department of the Katholieke Universiteit Leuven,\n
Belgium. Currently, Geert Leus is an Associate Professor at the Faculty of\n
Electrical Engineering, Mathematics and Computer Science of the Delft\n
University of Technology, The Netherlands. During the summer of 1998, he\n
visited Stanford University, and from March 2001 till May 2002 he was a\n
Visiting Researcher and Lecturer at the University of Minnesota. His\n
research interests are in the area of signal processing for communications.\n
Geert Leus received a 2002 IEEE Signal Processing Society Young Author Best\n
Paper Award and a 2005 IEEE Signal Processing Society Best Paper Award. He\n
is the Chair of the IEEE Signal Processing for Communications Technical\n
Committee, and an Associate Editor for the IEEE Transactions on Signal\n
Processing and the EURASIP Journal on Applied Signal Processing. In the\n
past, he has served on the Editorial Board of the IEEE Signal Processing\n
Letters and the IEEE Transactions on Wireless Communications.\n
Host: Prof. Urbashi Mitra, ubli@usc.edu, x0-4667
SEQUENCE:5
DTSTART:20101104T150000
LOCATION:EEB 248
DTSTAMP:20101104T150000
SUMMARY:Sparsity-Cognizant Total Least-Squares for Perturbed Compressive Sampling
UID:EC9439B1-FF65-11D6-9973-003065F99D04
DTEND:20101104T160000
END:VEVENT
END:VCALENDAR