BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:GenBench -- State-of-the-art generalisation research in NLP - Dieu
 wke Hupkes (Facebook AI Research\, ELLIS)
DTSTART:20230127T120000Z
DTEND:20230127T130000Z
UID:TALK196375@talks.cam.ac.uk
CONTACT:Michael Schlichtkrull
DESCRIPTION:Abstract:\n\nGood generalisation is of utmost importance for a
 ny artificial intelligence model. Traditionally\, the generalisation capab
 ilities of machine learning models are evaluated using random train/test s
 plits. However\, numerous recent studies have exposed substantial generali
 sation failures in models that perform well on such random within-distribu
 tion splits. So\, if random splitting is not good for measuring how robust
 ly models generalise to different scenarios\, how should we evaluate that?
  In this talk\, I present a taxonomy for characterising and understanding 
 generalisation in NLP\,  and use it to analyse over 400 papers of the ACL 
 anthology.\n\nBio:\n\nDieuwke Hupkes is a research scientist at FAIR. Prev
 iously\, she was a post-doctoral researcher at the University of Amsterdam
 \, where she also did her PhD. In her research\, she studies neural models
  of language processing\, in which she tries to incorporate knowledge from
  linguistics and philosophy of language. She is particularly excited about
  what neural models might teach us about language and human language proce
 ssing. \n\nTopic: NLIP Seminar\nTime: Jan 27\, 2023 12:00 PM London\n\nJoi
 n Zoom Meeting\nhttps://cl-cam-ac-uk.zoom.us/j/94330375053?pwd=TjRtbTg5aUd
 zWVdLRU15RjR0V2g0Zz09\n\nMeeting ID: 943 3037 5053\nPasscode: 768471\n
LOCATION:Virtual (Zoom)
END:VEVENT
END:VCALENDAR