BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Contextual dependencies in unsupervised word segmentation - Keith 
 Vertanen (University of Cambridge)
DTSTART:20070820T100000Z
DTEND:20070820T110000Z
UID:TALK7693@talks.cam.ac.uk
CONTACT:Philip Sterne
DESCRIPTION:We will be discussing the paper "Contextual dependencies in un
 supervised word segmentation" by Sharon Goldwater\, Thomas L. Griffiths an
 d Mark Johnson.\n \nAvailable from:\nhttp://cocosci.berkeley.edu/tom/paper
 s/wordseg1.pdf\n\nAbstract:\nDeveloping better methods for segmenting cont
 inuous text into words is important for improving the processing of Asian 
 languages\, and may shed light on how humans learn to segment speech. We p
 ropose two new Bayesian word segmentation methods that assume unigram and 
 bigram models of word dependencies respectively. The bigram model greatly 
 outperforms the unigram model (and previous probabilistic models)\, demons
 trating the importance of such dependencies for word segmentation. We also
  show that previous probabilistic models rely crucially on sub-optimal sea
 rch procedures. \n\n
LOCATION:TCM Seminar Room\, Cavendish Laboratory\, Department of Physics
END:VEVENT
END:VCALENDAR