BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:M-estimation strategies for the ranking problem - Nicolas Vayatis\
 , ENS Cachan
DTSTART:20110311T160000Z
DTEND:20110311T170000Z
UID:TALK28565@talks.cam.ac.uk
CONTACT:Richard Nickl
DESCRIPTION: Statistical learning theory was mainly developed in the frame
 work\nof binary classification under the assumption that observations in t
 he\ntraining set form an i.i.d. sample. The techniques involved in order t
 o\nprovide statistical guarantees for state-of-the-art learning algorithms
  are\nborrowed from the theory of empirical processes. This is made possib
 le not\nonly because of the "i.i.d." assumption on the data but also becau
 se of the\nnature of the performance measures\, such as classification err
 or or margin\nerror\, which are statistics of order one. In the talk\, I w
 ill  discuss a\nvariety of questions which arise in the theory when more i
 nvolved criteria\nare considered. The problem of bipartite ranking through
  ROC curve\noptimization provides a prolific source of optimization functi
 onals which\nare statistics of order strictly larger than one and several 
 examples will\nbe presented.\n\nhttp://www.cmla.ens-cachan.fr/Membres/vaya
 tis.html
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR