BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian Nonparametrics: Latent Feature and Prediction Models\, an
 d Efficient Inference - Piyush Rai (University of Utah)
DTSTART:20111011T103000Z
DTEND:20111011T113000Z
UID:TALK33298@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:Nonparametric Bayesian approaches offer a flexible modeling pa
 radigm for data without limiting the model-complexity a priori. The flexib
 ility comes from the fact that the model-complexity can grow adaptively wi
 th data. The Indian Buffet Process (IBP) is an example of a nonparametric 
 Bayesian model in which a set of observations are assumed to be generated 
 from a small set of latent features\, and the number of latent features ne
 ed not be known a priori. In this talk\, I will describe some of my recent
  work on the IBP based models\; in particular\, (1) A variant of the IBP w
 hich removes the independent latent features assumption\, and allows the l
 atent features to be related via a hierarchy\, (2) A nonparametric Bayesia
 n multitask learning model which uses a combination of the Dirichlet Proce
 ss mixture model and the IBP as the prior distribution on the weight vecto
 rs of multiple tasks\, and (3) An efficient\, search-based inference metho
 d for finding an approximate MAP estimate of the latent feature assignment
  matrix in the IBP based models.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR