BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Annotating Genericity: How Do Humans Decide?- A Case Study in Onto
 logy Extraction - Aurelie Herbelot\, Computer Laboratory\, University of C
 ambridge
DTSTART:20080125T120000Z
DTEND:20080125T130000Z
UID:TALK10367@talks.cam.ac.uk
CONTACT:Johanna Geiss
DESCRIPTION:This talk deals with the identification of kind versus non-kin
 d entities in natural language text for ontology extraction. The following
  two sentences illustrate the relevance of obtaining genericity annotation
 s for the creation of ontologies. -- the whale is a mammal -- the whale re
 scued the scuba diver. Given this input\, an ontology extraction system wo
 uld typically output the relationships 'whale -- is_a -- mammal' and 'whal
 e -- rescue -- scuba diver'. When inserted as such in an real-world ontolo
 gy\, these relations may give the user the false impression that 'one gene
 ral feature of whales is that they rescue scuba divers.' In order to preve
 nt this reading\, it is necessary to tag the first whale with a generic la
 bel and the second with a specific label.\n\nThe task of genericity annota
 tion using machine learning relies on a training corpus. Available corpora
 \, however\, are limited in the genres they cover and more importantly in 
 the range of labels that they use to describe the genericity phenomenon. T
 he public annotation schemes linked to those corpora are also often simpli
 fied and/or domain-specific. With the view of producing our own training c
 orpus\, we propose here an annotation scheme that covers the kind versus o
 bject distinction\, the specificity phenomenon and reference resolution. T
 he scheme is not domain-specific and produced\, over a small test set from
  the British National Corpus\, an inter-annotator agreement of Kappa = 0.7
 4.\n\nWe will discuss the scheme\, our choice of labels\, and the various 
 problems attached to the manual annotation of genericity. In particular\, 
 we will show the importance of reference resolution for accurate annotatio
 n.
LOCATION:SW01 Computer Laboratory
END:VEVENT
END:VCALENDAR