BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Recursive Deep Learning for Modeling Semantic Compositionality - R
 ichard Socher -  Stanford University
DTSTART:20130731T130000Z
DTEND:20130731T140000Z
UID:TALK46335@talks.cam.ac.uk
CONTACT:Tamara Polajnar
DESCRIPTION:Compositional and recursive structure is commonly found in dif
 ferent\nmodalities\, including natural language sentences and scene images
 . I\nwill introduce several recursive deep learning models that\, unlike s
 tandard deep learning methods can learn compositional meaning vector repre
 sentations for phrases\, sentences and images. These recursive neural netw
 ork based models obtain state-of-the-art performance on a variety of synta
 ctic and semantic language tasks such as parsing\, paraphrase detection\, 
 relation classification and sentiment analysis.\n\nBesides the good perfor
 mance\, the models capture interesting phenomena in language such as compo
 sitionality. For instance the models learn different types of high level n
 egation and how it can change the meaning of longer phrases with many posi
 tive words. They can learn that the sentiment following a "but" usually do
 minates that of phrases preceding the "but." Furthermore\, unlike many oth
 er machine learning approaches that rely on human designed feature sets\, 
 features are learned as part of the model.\n
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR