BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Convex Variational Bayesian Inference for Large Scale Generalized 
 Linear Models - Hannes Nickisch
DTSTART:20090827T100000Z
DTEND:20090827T110000Z
UID:TALK19664@talks.cam.ac.uk
CONTACT:Carl Edward Rasmussen
DESCRIPTION:Bayesian inference for most generalized linear models is analy
 tically not tractable. We show that a well known variational relaxation le
 ads to a convex problem for any log-concave model and provide a generic do
 uble loop algorithm for solving it on models with arbitrary super-Gaussian
  potentials. We iteratively decouple the criterion\, so that most of the c
 omputational work is done by solving large linear systems\, rendering our 
 algorithm much faster than previously proposed solvers. We evaluate our me
 thod on problems of Bayesian active learning for large binary classificati
 on models.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR