BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Identifying Non-Explicit Citing Sentences for Citation-Based Summa
 rization - Awais Athar (University of Cambridge)
DTSTART:20110203T120000Z
DTEND:20110203T130000Z
UID:TALK29733@talks.cam.ac.uk
CONTACT:Jimme Jardine
DESCRIPTION:http://www.aclweb.org/anthology/P/P10/P10-1057.pdf\n\nIdentify
 ing background (context) information in scientific articles can help scho
 lars understand major contributions in their\nresearch area more easily. I
 n this paper\,\nwe propose a general framework based\non probabilistic inf
 erence to extract such\ncontext information from scientific papers.\nWe m
 odel the sentences in an article and\ntheir lexical similarities as a Mark
 ov Random Field tuned to detect the patterns that\ncontext data create\, a
 nd employ a Belief\nPropagation mechanism to detect likely\ncontext senten
 ces. We also address the\nproblem of generating surveys of scientific pap
 ers. Our experiments show greater\npyramid scores for surveys generated us
 ing such context information rather than\ncitation sentences alone
LOCATION:GS15\, Computer Laboratory
END:VEVENT
END:VCALENDAR