BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning multi-domain dialogues - Paweł Budzianowski\, CUED\, Uni
 versity of Cambridge
DTSTART:20190301T120000Z
DTEND:20190301T130000Z
UID:TALK114535@talks.cam.ac.uk
CONTACT:Andrew Caines
DESCRIPTION:Even though machine learning has become the major scene in dia
 logue research community\, the real breakthrough has been blocked by the s
 cale of data available. To address this fundamental obstacle\, we introduc
 e the Multi-Domain Wizard-of-Oz dataset (MultiWOZ)\, a fully-labeled colle
 ction of human-human written conversations spanning over multiple domains 
 and topics. First results on multi-domain belief tracking and end-to-end m
 odelling will be analysed along with discussing open problems.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR