BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Fast and Reliable Online Learning to Rank for Information Retrieva
 l - Katja Hofmann\, University of Amsterdam
DTSTART:20120924T090000Z
DTEND:20120924T100000Z
UID:TALK39326@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Online learning to rank for information retrieval (IR) holds p
 romise for allowing the development of "self-learning search engines" that
  can automatically adjust to their users. With the large amount of e.g.\, 
 click data that can be collected in web search settings\, such techniques 
 could enable highly scalable ranking optimization. However\, feedback obta
 ined from user interactions is noisy\, and developing approaches that can 
 learn from this feedback quickly and reliably is a major challenge.\n\nIn 
 this talk I will present my recent work\, which addresses the challenges p
 osed by learning from natural user interactions. First\, I will detail a n
 ew method\, called Probabilistic Interleave\, for inferring user preferenc
 es from users' clicks on search results. I show that this method allows un
 biased and fine-grained ranker comparison using noisy click data\, and tha
 t this is the first such method that allows the effective reuse of histori
 cal data (i.e.\, collected for previous comparisons) to infer information 
 about new rankers. Second\, I show that Probabilistic Interleave enables n
 ew online learning to rank approaches that can reuse historical interactio
 n data to speed up learning by several orders of magnitude\, especially un
 der high levels of noise in user feedback. I conclude with an outlook on r
 esearch directions in online learning to rank for IR\, that are opened up 
 by our results.\n
LOCATION:Small lecture theatre\, Microsoft Research Ltd\, 7 J J Thomson Av
 enue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR