BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Multiresolution Matrix Factorization - Prof Risi Kondor (U Chicago
 )
DTSTART:20160913T100000Z
DTEND:20160913T110000Z
UID:TALK67495@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:The sheer size of today's datasets dictates that learning algo
 rithms compress or reduce their input data and/or make use of parallelism.
  Multiresolution Matrix Factorization (MMF) makes a connection between suc
 h computational strategies and some classical themes in Applied Mathematic
 s\, namely Multiresolution Analysis and Multigrid Methods. In particular\,
  the similarity matrices appearing in data often have multiresolution stru
 cture\, which can be exploited both for learning and to facilitate computa
 tion. \nIn addition to the general MMF framework\, I will present our new\
 , high performance parallel MMF software library and show some results on 
 matrix compression/sketching tasks (joint work with Nedelina Teneva\, Pram
 od Mudrakarta and Vikas Garg).
LOCATION:CBL Room BE-438
END:VEVENT
END:VCALENDAR