BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Predictive Uncertainty in Deep Learning - Andrey Malinin\, CUED\, 
 University of Cambridge
DTSTART:20180504T110000Z
DTEND:20180504T120000Z
UID:TALK104473@talks.cam.ac.uk
CONTACT:Andrew Caines
DESCRIPTION:Estimating uncertainty is important to improving the safety of
  AI systems. Current methods do not explicitly model predictive uncertaint
 y arising from distributional mismatch between training and test data\, ei
 ther conflating  distributional uncertainty with data uncertainty or impli
 citly modeling it via model uncertainty. Thus\, a new framework for modeli
 ng predictive uncertainty\, called Prior Networks (PNs)\, is proposed. Pri
 or Networks allow each source of predictive uncertainty - data\, distribut
 ional and model uncertainty\, to be modeled explicitly. This work examines
  predictive uncertainty in the context of classification. Prior Networks a
 re evaluated on the tasks of identifying out-of-distribution (OOD) samples
  and detecting misclassification on the MNIST dataset\, where they are fou
 nd to outperform previous methods.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR