BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Talking with Reddit: Exploring limitations of recurrent neural net
 work conversational systems. - Alexander Gaunt - Trinity College\, Cambrid
 ge
DTSTART:20151008T090000Z
DTEND:20151008T100000Z
UID:TALK61643@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Recently\, there has been great interest in building conversat
 ional systems based on recurrent neural networks. If trained on a large-sc
 ale corpus of conversation-like data\, these models can produce surprising
 ly fluent responses when interacting with a human. I will discuss my imple
 mentation of a neural network conversational agent and describe the output
  of this model after training on Reddit conversations. While this model is
  able to produce plausible responses in a variety of settings\, I will con
 centrate on the cases where the model fails\, and briefly describe attempt
 s to incorporate information from knowledge-bases to mitigate these failur
 es.Abstract not available
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station Road\, Cambridge
 \, CB1 2FB
END:VEVENT
END:VCALENDAR