BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Monolingual and multilingual\, explicit and latent vector represen
 tations of meaning - Roberto Navigli (Sapienza University of Rome)
DTSTART:20161110T110000Z
DTEND:20161110T120000Z
UID:TALK68124@talks.cam.ac.uk
CONTACT:Mohammad Taher Pilehvar
DESCRIPTION:In this talk I will present different kinds of representation 
 of word senses and concepts. I will start with latent representations obta
 ined as *sense embeddings* from the application of word2vec to the Wikiped
 ia corpus\, sense-tagged with a multilingual disambiguation algorithm base
 d on *BabelNet*\, the largest multilingual semantic network and encycloped
 ic dictionary covering 14 million concepts and entities and 271 languages.
 \n\nI will then move on to two explicit vector representations of meaning 
 (*NASARI*)\, based on lexical co-occurrence and multilingual semantic gene
 ralization\, respectively\, and a third latent version obtained from the w
 ord embeddings of the lexical vector.\n\nExperimental results in several t
 asks\, including word similarity\, sense clustering\, identification of se
 nse predominance\, and word sense disambiguation highlight high performanc
 e and show that\, whenever a comparison is possible\, sense representation
 s consistently outperform word representations.\n\nThis is joint work with
  José Camacho-Collados\, Ignacio Iacobacci and Mohammad Taher Pilehvar.\n
LOCATION:GR05\, English Faculty\, 9 West Road (Sidgwick Site)
END:VEVENT
END:VCALENDAR