BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Citations and Argumentation for Better Information Access - Simone
  Teufel\, Computer Laboratory
DTSTART:20080514T131500Z
DTEND:20080514T141500Z
UID:TALK11958@talks.cam.ac.uk
CONTACT:Timothy G. Griffin
DESCRIPTION:When publishing a scientific paper\, authors pay great attenti
 on to how they\ntalk about previous work: They cite it (in a easy-to-detec
 t way)\, compare\ntheir own work to it (which is harder to detect)\, and m
 ake evaluative\nstatements about its quality (which is the hardest to dete
 ct). All of these\npieces of information can be invaluable when building i
 nformation access\nsystems such as summarisers or citation indexers. But w
 ithout NLP\, only\ncitations are readily detectable (and Google Scholar an
 d similar citation\nindexers do). My work looks at how to use discourse co
 ntext and NLP to tease\nmore information out of the text: sentiment toward
 s citations\, explicit\nstatements of self-praise and comparison to others
 . Automatic annotation is\nbased on supervised machine learning from lower
 -level sentential features. The\ndata used comes from two different domain
 s (computational linguistics and\nchemistry). Results are in the form of h
 uman annotation agreement\, similarity\nof automatic and human annotations
 \, and measures of usefulness for search.
LOCATION:Lecture Theatre 1\, Computer Laboratory
END:VEVENT
END:VCALENDAR