BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Coupling Semi-Supervised Learning of Categories and Relations - St
 uart Moore (University of Cambridge)
DTSTART:20091102T123000Z
DTEND:20091102T133000Z
UID:TALK21297@talks.cam.ac.uk
CONTACT:Diarmuid Ó Séaghdha
DESCRIPTION:At this session of the NLIP Reading Group we'll be discussing 
 the following paper:\n\nAndrew Carlson\, Justin Betteridge\, Estevam R. Hr
 uschka Jr. and Tom M. Mitchell. 2009. "Coupling semi-supervised learning o
 f categories and relations":http://aclweb.org/anthology-new/W/W09/W09-2201
 .pdf. In Proceedings of the NAACL-HLT-09 Workshop on Semi-Supervised Learn
 ing for Natural Language Processing.\n\n*Abstract:*\nWe consider semi-supe
 rvised learning of information extraction methods\, especially for extract
 ing instances of noun categories (e.g.\, 'athlete'\, 'team') and relations
  (e.g.\, 'playsForTeam(athlete\,team)'). Semi-supervised\napproaches using
  a small number of labeled examples together with many unlabeled examples 
 are often unreliable as they frequently produce an\ninternally consistent\
 , but nevertheless incorrect set of extractions. We propose that this prob
 lem can be overcome by simultaneously learning classifiers for many differ
 ent categories and relations in the presence of an ontology defining const
 raints that couple the training of these classifiers.\nExperimental result
 s show that simultaneously learning a coupled collection of classifiers fo
 r 30 categories and relations results in much more accurate\nextractions t
 han training classifiers individually.
LOCATION:GS15\, Computer Laboratory
END:VEVENT
END:VCALENDAR