BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Conversation Trees: A Grammar Model for Topic Structure in Online 
 Forums - Annie Louis\, University of Essex
DTSTART:20160520T110000Z
DTEND:20160520T120000Z
UID:TALK65228@talks.cam.ac.uk
CONTACT:Kris Cao
DESCRIPTION:Online forums are commonplace today and used for various purpo
 ses: product support and troubleshooting\, opining about events and people
 \, and student interaction on online course platforms. Threads in these fo
 rums become long\, involve posts from multiple users\, and the chronologic
 al order of the posts in a thread does not represent a continuous flow of 
 dialog. Adding structure to these threads is important for tasks such as i
 nformation extraction\, search\, and summarization. One such aspect of str
 ucture is topic: a discussion often involves a number of subtopics. In thi
 s talk\, I will present work aiming to characterize the content of a forum
  thread as a conversation tree of topics.\n\nOur models jointly perform tw
 o tasks: (i) segment a thread into sub-parts\, and (ii) assign a topic to 
 each part. The key idea in this work is a formalization of topic structure
  using probabilistic grammars. By leveraging the flexibility of two gramma
 r formalisms\, Context-Free Grammars and Linear Context-Free Rewriting Sys
 tems\, our models create desirable structures for forum threads: our topic
  segmentation is hierarchical\, links non-adjacent segments on the same to
 pic\, and jointly labels the topic during segmentation. I will show that t
 hese models outperform a number of tree generation baselines.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR