BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Multidimensional Dialogue Management and Dialogue Act Recognition 
 using Bayesian Networks - Simon Keizer (University of Tilburg)
DTSTART:20070814T103000Z
DTEND:20070814T113000Z
UID:TALK7809@talks.cam.ac.uk
CONTACT:Dr Marcus Tomalin
DESCRIPTION:The first part of the talk I will present my current work as a
  postdoc\nresearcher on the PARADIME project (PARallel Agent-based DIalogu
 e Management\nEngine).  In the project\, I have developed a dialogue manag
 er (DM) that takes\nthe multidimensional nature of communication into acco
 unt.  The DM supports\nthe generation of multifunctional system utterances
  via dialogue act\ncontributions from several agents\, that each address o
 ne particular dimension\nof communication.  The DM is integrated in the de
 monstration system for\ninteractive question answering as developed within
  the Dutch research program\nIMIX (Interactive Multimodal Information eXtr
 action).\n\nBecause of its particular relevance to the research on probabi
 listic dialogue\nmodelling in Cambridge\, the second part of the talk will
  be about my PhD\nresearch on using Bayesian Networks (BNs) for dialogue m
 odelling.  This work\nparticularly consisted of experiments on using BNs f
 or dialogue act\nrecognition\, involving the classification of Forward- an
 d Backward-looking\nfunctions from a DAMSL based annotation scheme.
LOCATION:LR5\, Engineering Department\, Baker Building
END:VEVENT
END:VCALENDAR