BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Alternating proximal gradient descent for nonconvex regularised pr
 oblems with multiconvex coupling terms - Mila Nikolova (CNRS (Centre natio
 nal de la recherche scientifique)\; ENS de Cachan)
DTSTART:20170908T080000Z
DTEND:20170908T085000Z
UID:TALK78451@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Co-author: Pauline Tan <br><br>There has been an increasing in
 terest in constrained nonconvex&nbsp\; regularized block multiconvex optim
 ization problems. We introduce an&nbsp\; approach that effectively exploit
 s the multiconvex structure of the coupling term and enables complex appli
 cation-dependent regularization terms to be used. The proposed Alternating
  Structure-Adapted Proximal gradient descent algorithm enjoys simple well 
 defined updates. Global convergence of the algorithm to a critical point i
 s proved using the so-called Kurdyka-Lojasiewicz&nbsp\; property. What is 
 more\, we prove that a large class of useful objective functions obeying o
 ur assumptions are subanalytic and thus satisfy the Kurdyka-Lojasiewicz pr
 operty. Finally\, present an application of the algorithm to big-data air-
 born sequences of images.<br><br><br><br>
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR