BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Combining Manual Rules and Supervised Learning for Hedge Cue and S
 cope Detection - Marek Rei\, University of Cambridge
DTSTART:20100625T110000Z
DTEND:20100625T113000Z
UID:TALK25382@talks.cam.ac.uk
CONTACT:Laura Rimell
DESCRIPTION:Hedge cues were detected using a supervised Conditional Random
  Field\n(CRF) classifier exploiting features from the RASP parser. The CR
 F’s\npredictions were filtered using known cues and unseen instances we
 re\nremoved\, increasing precision while retaining recall. Rules for scope
 \ndetection\, based on the grammatical relations of the sentence and the\n
 part-of-speech tag of the cue\, were manually developed. However\, another
 \nsupervised CRF classifier was used to refine these predictions. As a 
 final\nstep\, scopes were constructed from the classifier output using a
  small\nset of post-processing rules. Development of the system revealed a
 \nnumber of issues with the annotation scheme adopted by the organisers.\n
 \nJoint work with Ted Briscoe.
LOCATION:GS15\, Computer Laboratory
END:VEVENT
END:VCALENDAR