BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Gauge Equivariant Convolutional Networks on Manifolds - Taco Cohen
DTSTART:20190131T133000Z
DTEND:20190131T143000Z
UID:TALK119401@talks.cam.ac.uk
CONTACT:Adrian Weller
DESCRIPTION:Equivariance to symmetry transformations is one of the first r
 ational principles for neural network architecture design. Equivariant net
 works have shown excellent performance on vision and medical imaging probl
 ems that exhibit symmetries. In this talk\, I will show how this principle
  can be extended to data defined on general manifolds\, using ideas from t
 heoretical physics. We use the new theory to develop a highly practical an
 d scalable alternative to Spherical CNNs\, and show that this method outpe
 rforms previous methods on global climate pattern segmentation and omnidir
 ectional vision.
LOCATION:Engineering Department\, CBL Room BE-438.
END:VEVENT
END:VCALENDAR