BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Optimization meets Statistics: Fast global convergence for high-di
 mensional statistical recovery - Martin Wainwright\, University of Califor
 nia\, Berkeley
DTSTART:20130222T160000Z
DTEND:20130222T170000Z
UID:TALK42546@talks.cam.ac.uk
CONTACT:Richard Samworth
DESCRIPTION:Many methods for solving high-dimensional statistical\ninverse
  problems are based on convex optimization problems formed by\nthe weighte
 d sum of a loss function with a norm-based regularizer.\nParticular exampl
 es include $\\ell_1$-based methods for sparse vectors\nand matrices\, nucl
 ear norm for low-rank matrices\, and various\ncombinations thereof for mat
 rix decomposition and robust PCA.  In this\ntalk\, we describe an interest
 ing connection between computational and\nstatistical efficiency\, in part
 icular showing that the same conditions\nthat guarantee that an estimator 
 has good statistical error can also\nbe used to certify fast convergence o
 f first-order optimization\nmethods up to statistical precision.
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR