BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Graph-Guided Banding for Covariance Estimation - Jacob Bien (Corne
 ll University)
DTSTART:20151211T143000Z
DTEND:20151211T153000Z
UID:TALK60807@talks.cam.ac.uk
CONTACT:Quentin Berthet
DESCRIPTION:Reliable estimation of the covariance matrix is notoriously di
 fficult in high dimensions. Numerous methods assume that the population co
 variance (or inverse covariance) matrix is sparse while making no particul
 ar structural assumptions on the desired sparsity pattern. A highly-relate
 d\, yet complementary\, literature studies the setting in which the measur
 ed variables have a known ordering\, in which case a banded (or near-bande
 d) population matrix is assumed. This work focuses on the broad middle gro
 und that lies between the former approach of complete neutrality to the sp
 arsity pattern and the latter highly restrictive assumption of having a kn
 own ordering. We develop a class of convex regularizers that is in the spi
 rit of banding and yet attains sparsity structures that can be customized 
 to a wide variety of applications.
LOCATION:MR12\, Centre for Mathematical Sciences\, Wilberforce Road\, Camb
 ridge.
END:VEVENT
END:VCALENDAR