BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Modelling syntactico-semantic composition for natural language und
 erstanding and generation - Dr Weiwei Sun - Department of Computer Science
  and Technology\, University of Cambridge
DTSTART:20210310T150000Z
DTEND:20210310T160000Z
UID:TALK156910@talks.cam.ac.uk
CONTACT:Ben Karniely
DESCRIPTION:One of the central problems in language technology is the real
 ization of an accurate mapping between natural language utterances and in-
 depth meaning representations. In this talk\, I will discuss methods for t
 his bi-directional mapping which exploit graph-centric representations\, f
 ormalisms\, algorithms and neural networks. I will first introduce a neura
 l graph rewriting framework to model syntactico-semantic composition which
  combines the strengths of Hyperedge Replacement Grammar (HRG)\, for knowl
 edge-intensive learning\, and Graph Neural Networks\, for data-intensive l
 earning. I will then discuss two fundamental problems in this framework: s
 emantic graph parsing and parsing semantic graphs\, i.e. computing all pos
 sible / the best derivation(s) of a given string or graph. In particular\,
  I will demonstrate that exact graph parsing can be efficient for large gr
 aphs and with large grammars. With the ability to enumerate every derivati
 on of a surface string or a semantic graph\,  we are ready to build practi
 cal language understanding and generation systems. I will report on neural
  systems that achieve state-of-the-art accuracy for English Resource Seman
 tics. I will conclude by laying out a few ideas for future work on meaning
  representation--mediated Natural Language Processing.\n\n\nLink to join: 
 https://cl-cam-ac-uk.zoom.us/j/91253900399?pwd=SU5TNnpYdDlQbzQ4SEVPVWVWa0N
 ldz09\n\nA recording of this talk is available to members of the Departmen
 t at the following link: https://www.cl.cam.ac.uk/seminars/wednesday/video
 /
LOCATION:Online
END:VEVENT
END:VCALENDAR