BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Urban Dictionary Embeddings for Slang NLP Applications - Dr Barbar
 a McGillivray (University of Cambridge and The Alan Turing Institute)
DTSTART:20200130T110000Z
DTEND:20200130T120000Z
UID:TALK138889@talks.cam.ac.uk
CONTACT:Qianchu Liu
DESCRIPTION:The choice of the corpus on which word embeddings are trained 
 can have a sizable effect on the learned representations\, the types of an
 alyses that can be performed with them\, and their utility as features for
  machine learning models. In this talk I will present my work on the first
  set of word embeddings trained on the content of Urban Dictionary\, a cro
 wd-sourced dictionary for slang words and phrases. I will show that althou
 gh these embeddings are trained on fewer total tokens\, they have high per
 formance across a range of common word embedding evaluations\, ranging fro
 m semantic similarity to word clustering tasks. Further\, for some extrins
 ic tasks such as sentiment analysis and sarcasm detection where we expect 
 to require some knowledge of colloquial language on social media data\, in
 itializing classifiers with the Urban Dictionary Embeddings resulted in im
 proved performance compared to initializing with a range of other well-kno
 wn\, pre-trained embeddings that are order of magnitude larger in size.
LOCATION:GR04\, Faculty of English\, 9 West Rd (Sidgwick Site)
END:VEVENT
END:VCALENDAR