BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Variational Methods and Compressed Sensing - Nilesh Tripuraneni (U
 niversity of Cambridge)
DTSTART:20160407T133000Z
DTEND:20160407T150000Z
UID:TALK65400@talks.cam.ac.uk
CONTACT:Yingzhen Li
DESCRIPTION:Our focus will be to first review the compressed sensing recon
 struction problem and cast it in the language of probabilistic inference. 
 We will then show how approximate inference methods (mean-field variationa
 l inference/loopy belief propagation) can be used to derive large-scale so
 lvers for compressed sensing reconstruction which closely resemble proxima
 l gradient algorithms (and have connections to statistical physics). Along
  the way we will also explore how different forms of sparsity (l0-like spi
 ke-and-slab priors) can be naturally be incorporated into the probabilisti
 c framework and their effect on the sparse reconstruction phase transition
 .\n\nThe talk will not assume prior background about compressed sensing bu
 t familiarity with graphical models and approximate inference (loopy belie
 f propagation/variational inference) will be helpful. The talk will also b
 e conducted using overhead slides\, with a few calculations performed in d
 etail\, so having paper/pencil handy will be useful to follow along. 
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR