BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Convolutional Neural Networks on Graphs - Xavier Bresson (Nanyang 
 Technological University)
DTSTART:20171101T095000Z
DTEND:20171101T104000Z
UID:TALK94252@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Convolutional neural networks have greatly improved state-of-t
 he-art performances in computer vision and speech analysis tasks\, due to 
 its high ability to extract multiple levels of representations of data. In
  this talk\, we are interested in generalizing convolutional neural networ
 ks from low-dimensional regular grids\, where image\, video and speech are
  represented\, to high-dimensional irregular domains\, such as social netw
 orks\, telecommunication networks\, or words&#39\; embedding. We present a
  formulation of convolutional neural networks on graphs in the context of 
 spectral graph theory\, which provides the necessary mathematical backgrou
 nd and efficient numerical schemes to design fast localized convolutional 
 filters on graphs. Numerical experiments demonstrate the ability of the sy
 stem to learn local stationary features on graphs.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR