BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Structural Expectation Propagation (SEP): Bayesian structure learn
 ing for networks with latent variables  - Nevena Lazic (Microsoft Research
  Cambridge)
DTSTART:20130130T110000Z
DTEND:20130130T120000Z
UID:TALK43305@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:Learning the structure of discrete Bayesian networks has been 
 the subject of extensive research in machine learning.  One of the few met
 hods that can handle networks with latent variables is the "structural EM 
 algorithm"  which interleaves greedy structure search with the estimation 
 of latent variables and parameters\, maintaining a single best network at 
 each step.\nI will describe Structural Expectation Propagation (SEP)\, a n
 ovel method that iteratively updates a variational posterior distribution 
 over structure\, latent variables\, and parameters.  Conventional EP makes
  local distribution updates based on a 'context' that captures uncertainty
  in the remainder of the network. SEP extends this context to include unce
 rtainty in structure\, and returns a variational distribution over structu
 res rather than a single network. \nI will demonstrate the performance of 
 SEP on synthetic problems\, as well as on real-world data coming from a cl
 inical study of asthma and allergies. \n
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR