BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Machine Learning @ CUED
SUMMARY:Variational autoencoders with latent graphical mod
 els - Prof David Duvenaud (University of Toronto)
DTSTART;TZID=Europe/London:20161213T110000
DTEND;TZID=Europe/London:20161213T120000
UID:TALK67586AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/67586
DESCRIPTION:We propose a general modeling and inference framew
 ork that composes probabilistic graphical models w
 ith deep learning methods\, in a way that combines
  their respective strengths. Our model family comb
 ines graphical structure in latent variables with 
 neural network observation models. For inference\,
  we use variational recognition networks to produc
 e  local evidence summaries\, and combine them usi
 ng exact graphical model inference.  We illustrate
  this framework with several example models\, and 
 an application to automatic mouse behavior modelin
 g from video.
LOCATION:CBL Room BE-438\, Department of Engineering
CONTACT:Zoubin Ghahramani
END:VEVENT
END:VCALENDAR