BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:PAC learning - Peter Rugg\, Churchill College
DTSTART:20161012T180000Z
DTEND:20161012T184000Z
UID:TALK67735@talks.cam.ac.uk
CONTACT:Jasper Lee
DESCRIPTION:Machine learning allows computers to solve problems without be
 ing explicitly programmed with the solution. However\, what sorts of probl
 ems can be learned? What does it even mean to learn a problem?\nThe Probab
 ly Approximately Correct (PAC) machine learning framework addresses these 
 questions\, specifying worst case error bounds before a problem can be sai
 d to be learnable. In this talk\, we will see a formulation of the supervi
 sed binary classification problem\, followed by a definition of PAC learni
 ng. We will then try to find a way of determining which problems are PAC l
 earnable\, and which are not.
LOCATION:Wolfson Hall\, Churchill College
END:VEVENT
END:VCALENDAR