BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Generalization Bounds via Online Learning - Dr Gergely Neu\, Unive
 rsitat Pompeu Fabra
DTSTART:20230308T140000Z
DTEND:20230308T150000Z
UID:TALK194128@talks.cam.ac.uk
CONTACT:Prof. Ramji Venkataramanan
DESCRIPTION:Bounding the generalization error is one most fundamental\npro
 blems in statistical learning theory. In this talk\, I will present\na new
  framework for deriving generalization bounds from the\nperspective of onl
 ine learning. Specifically\, we construct an online\nlearning game called 
 the Generalization Game\, where an online learner\nis trying to compete wi
 th a fixed statistical learning algorithm in\npredicting the sequence of g
 eneralization gaps on a training set of\ni.i.d. data points. As I will sho
 w\, this framework will allow us to\nrecover a range of classic bounds inc
 luding PAC-Bayes and\ngeneralizations thereof. (Based on joint work with G
 abor Lugosi.)
LOCATION:MR5\, CMS Pavilion A
END:VEVENT
END:VCALENDAR