BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:CANCELLED: Local Deep Kernel Learning for Efficient Non-linear SVM
  Prediction - Manik Varma (Microsoft Research India)
DTSTART:20130927T100000Z
DTEND:20130927T110000Z
UID:TALK47307@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:THIS TALK HAS BEEN CANCELLED:\nOur objective is to speed up no
 n-linear SVM prediction while maintaining classification accuracy above an
  acceptable limit. We generalize Localized Multiple Kernel Learning so as 
 to learn a tree-based primal feature embedding which is high dimensional a
 nd sparse. Primal based classification decouples prediction costs from the
  number of support vectors and our tree-structured features efficiently en
 code non-linearities while speeding up prediction exponentially over the s
 tate-of-the-art. We develop routines for optimizing over the space of tree
 -structured features and efficiently scale to problems with more than half
  a million training points. Experiments on benchmark data sets reveal that
  our formulation can reduce prediction costs by more than three orders of 
 magnitude in some cases with a moderate sacrifice in classification accura
 cy as compared to RBF-SVMs. Furthermore\, our formulation leads to better 
 classification accuracies over leading methods.
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR