BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Inducing Meaning from Text - Dan Jurafsky (Stanford University)
DTSTART:20080506T120000Z
DTEND:20080506T130000Z
UID:TALK12019@talks.cam.ac.uk
CONTACT:Dr Marcus Tomalin
DESCRIPTION:Online models of word meaning (like dictionaries and thesauri)
  or\nworld knowledge (like scripts or narratives) are crucial for natural\
 nlanguage understanding.  Could we learn these meanings automatically\nfro
 m text?  I first report on joint work with Rion Snow and Andrew\nNg on ind
 ucing the meaning of words from text on the Web in the\ncontext of augment
 ing WordNet\, a large online thesaurus of English.\nThese include a semi-s
 upervised method for learning when a new word\nis a `hypernym' or in the '
 is-a' relation with another word\, a new\nprobabilistic algorithm for comb
 ining evidence from multiple relation\ndetectors\, and a algorithm for clu
 stering the induced word senses.\nI then report on joint work with Nate Ch
 ambers on inducing `narratives'\,\na script-like sequence of events that f
 ollow a protagonist.  This\nwork includes inducing the relations between e
 vents\, ordering the\nrelations and clustering them into prototype narrati
 ves.
LOCATION:LR5\, Engineering Department\, Baker Building
END:VEVENT
END:VCALENDAR