BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Calibrating prediction uncertainty with validity guarantees - Vlad
 imir Vovk (Royal Holloway\, University of London)
DTSTART:20250602T104500Z
DTEND:20250602T114500Z
UID:TALK230575@talks.cam.ac.uk
DESCRIPTION:My plan is to review several related methods for calibrating p
 rediction uncertainty that enjoy various properties of validity under assu
 mptions that are standard in mainstream machine learning. The method with 
 the simplest validity guarantees is conformal prediction\, which produces 
 prediction sets with a given bound on the probability of error. Venn predi
 ction and conformal predictive distributions are methods of probabilistic 
 prediction that work in the case of classification and regression\, respec
 tively\; their validity guarantees are more complicated but still natural.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR