BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:How much linguistics is needed for NLP? - Edward Grefenstette
DTSTART:20151106T120000Z
DTEND:20151106T130000Z
UID:TALK61952@talks.cam.ac.uk
CONTACT:Kris Cao
DESCRIPTION:Many problems in Natural Language Processing\, from Machine Tr
 anslation to Parsing\, can be viewed as transduction tasks. Recently\, seq
 uence-to-sequence mapping approaches using recurrent networks and parallel
  corpora have shown themselves to be capable of learning fairly complex tr
 ansductions without the need for heavy (or any) annotation or alignment da
 ta. Traditional linguistically-motivated features such as syntactic types 
 and dependencies are entirely latent in such models\, reducing the need fo
 r expert linguistic knowledge in designing new solutions in NLP. In this t
 alk\, I will discuss the strengths and weaknesses of such approaches\, bef
 ore presenting some ameliorations based on attention mechanisms and workin
 g memory enhancements to standard recurrent neural networks.
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR