BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Imitation learning for structured prediction and automated fact ch
 ecking - Dr Andreas Vlachos\, University of Sheffield
DTSTART:20171124T120000Z
DTEND:20171124T130000Z
UID:TALK82231@talks.cam.ac.uk
CONTACT:Anita Verő
DESCRIPTION:In the first part of this talk\, I will introduce our work on 
 imitation learning\, a learning paradigm originally developed to learn rob
 otic controllers from demonstrations by humans\, e.g. autonomous helicopte
 rs from pilot's demonstrations. Recently\, algorithms for structure predic
 tion were proposed under this paradigm and have been applied successfully 
 to a number of tasks such as dependency parsing\, information extraction\,
  coreference resolution and semantic parsing. Key advantages are the abili
 ty to handle large output search spaces and to learn with non-decomposable
  loss functions. In this talk I will give a detailed overview of imitation
  leaning\, discuss its relation to other learning paradigms\, describe som
 e recent applications\, including natural language generation\, abstract m
 eaning representation parsing and its use in training recurrent neural net
 works.\nIn the second part of this talk\, I will give an overview of our w
 ork on automated fact checking\, and how it relates to the wider efforts o
 n battling misinformation and rumour detection.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR