BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:&quot\;Structured sparsity and convex optimization&quot\; - Franci
 s Bach
DTSTART:20120503T130000Z
DTEND:20120503T143000Z
UID:TALK37507@talks.cam.ac.uk
CONTACT:Konstantina Palla
DESCRIPTION: The concept of parsimony is central in many scientific\n doma
 ins. In the context of statistics\, signal processing or machine\n learnin
 g\, it takes the form of variable or feature selection problems\,\n and is
  commonly used in two situations: First\,  to make the model or\n the pred
 iction more interpretable or cheaper to use\, i.e.\, even if the\n underly
 ing problem does not admit sparse solutions\, one looks for the\n best spa
 rse approximation. Second\, sparsity can also be used given\n prior knowle
 dge that the model should be sparse. In these two\n situations\, reducing 
 parsimony to finding models with low cardinality\n turns out to be limitin
 g\, and structured parsimony has emerged as a\n fruitful practical extensi
 on\, with applications to image processing\,\n text processing or bioinfor
 matics. In this talk\, I will review recent\n results on structured sparsi
 ty\, as it applies to machine learning and\n signal processing. (joint wor
 k with R. Jenatton\, J. Mairal and G.\n Obozinski)
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR