BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A New Corpus and Imitation Learning Framework for Context-Dependen
 t Semantic Parsing - Andreas Vlachos\, UCL
DTSTART:20150116T120000Z
DTEND:20150116T130000Z
UID:TALK56067@talks.cam.ac.uk
CONTACT:Tamara Polajnar
DESCRIPTION:Semantic parsing is the task of translating natural language u
 tterances into a machine-interpretable meaning representation. Most approa
 ches to this task have been evaluated on a small number of existing corpor
 a which assume that all utterances must be interpreted according to a data
 base and typically ignore context. In this paper we present a new\, public
 ly available corpus for context-dependent semantic parsing. The MRL used f
 or the annotation was designed to support a portable\, interactive tourist
  information system. We develop a semantic parser for this corpus by adapt
 ing the imitation learning algorithm DAGGER without requiring alignment in
 formation during training. DAGGER improves upon independently trained clas
 sifiers by 9.0 and 4.8 points in F-score on the development and test sets 
 respectively.\n\nJoint work with Stephen Clark.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR