BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:NLP Reading Group: Learning Continuous Phrase Representations and 
 Syntactic Parsing with Recursive Neural Networks - Wenduan Xu
DTSTART:20120315T120000Z
DTEND:20120315T130000Z
UID:TALK36960@talks.cam.ac.uk
CONTACT:Marek Rei
DESCRIPTION:This week Wenduan will be presenting:\n\nLearning Continuous P
 hrase Representations and Syntactic Parsing with Recursive Neural Networks
 . \nRichard Socher\, Christopher D. Manning\, Andrew Y. Ng. \nDeep Learnin
 g and Unsupervised Feature Learning Workshop - NIPS 2010\n\nNatural langua
 ge parsing has typically been done with small sets of discrete categories 
 such as NP and VP\, but this representation does not capture the full synt
 actic nor semantic richness of linguistic phrases\, and attempts to improv
 e on this by lexicalizing phrases only partly address the problem at the c
 ost of huge feature spaces and sparseness. To address this\, we introduce 
 a recursive neural network architecture for jointly parsing natural langua
 ge and learning vector space representations for variable-sized inputs. At
  the core of our architecture are context-sensitive recursive neural netwo
 rks (CRNN). These networks can induce distributed feature representations 
 for unseen phrases and provide syntactic information to accurately predict
  phrase structure trees. Most excitingly\, the representation of each phra
 se also captures semantic information: For instance\, the phrases “decli
 ne to comment” and “would not disclose the terms” are close by in th
 e induced embedding space. Our current system achieves an unlabeled bracke
 ting F-measure of 92.1% on the Wall Street Journal dataset for sentences u
 p to length 15.\n\nPDF: http://www.cs.stanford.edu/people/ang/papers/nipsd
 lufl10-LearningContinuousPhraseRepresentations.pdf\n\nMarch 15\, 12:00\, G
 S15\n
LOCATION:GS15\, Computer Laboratory
END:VEVENT
END:VCALENDAR