BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning to Learn for Structured Sparsity - Nino Shervashidze (INR
 IA)
DTSTART:20140311T110000Z
DTEND:20140311T120000Z
UID:TALK51055@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:Structured sparsity has recently emerged in statistics\, machi
 ne learning and signal processing as a promising paradigm for learning in 
 high-dimensional settings. A number of methods have been proposed for lear
 ning under the assumption of structured sparsity\, including group LASSO a
 nd graph LASSO. All of these methods rely on prior knowledge on how to pen
 alize individual subsets of variables during the subset selection process.
  However\, these weights on groups of variables are in general unknown. In
 ferring group weights from data is a key open problem in research on struc
 tured sparsity.\n\nIn this work\, we propose a probabilistic approach to t
 he problem of group weight learning. We model the group weights as hyperpa
 rameters of heavy-tailed priors on groups of variables and derive an appro
 ximate inference scheme to infer these hyperparameters. We empirically sho
 w that we are able to recover the model hyperparameters when the data are 
 generated from the model and demonstrate the utility of learning group wei
 ghts in denoising problems.
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR