BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Unsupervised Domain Tuning to Improve Word Sense Disambiguation - 
 Judita Preiss\, University of Sheffield
DTSTART:20130315T120000Z
DTEND:20130315T130000Z
UID:TALK43217@talks.cam.ac.uk
CONTACT:Ekaterina Kochmar
DESCRIPTION:The topic of a document can prove to be useful information for
  word sense disambiguation since certain meanings tend to be associated wi
 th particular topics. I will discuss three techniques which improve all ex
 isting WSD algorithms we tested by establishing a sense per (Latent Dirich
 let allocation based) topic\, and do not require any modification to the u
 nderlying algorithm to yield a performance increase. The evaluation of thr
 ee unsupervised and one supervised algorithms is carried out on the sport 
 and finance domains.\n\nJoint work with Mark Stevenson.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR