BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Expectations vs. Reality: Lessons learned from Working on Toxic Co
 ntent Detection in NLP - Nedjma Ousidhoum (University of Cambridge)
DTSTART:20220121T120000Z
DTEND:20220121T130000Z
UID:TALK168302@talks.cam.ac.uk
CONTACT:Michael Schlichtkrull
DESCRIPTION:Join Zoom Meeting https://cl-cam-ac-uk.zoom.us/j/99831805544?p
 wd=NUMrTGE4K2U3V2h0NlhtTHNsOG5rQT09\n\nMeeting ID: 998 3180 5544 Passcode:
  779252\n\nIn order to improve the online moderation process\, there has b
 een an increasing need for building toxic language detection tools that do
  not only flag bad words\, but rather filter out toxic content in a more n
 uanced fashion. In order to train such models\, it is essential to acquire
  data of high quality. However\, in the absence of universal definitions o
 f terms such as hate speech\, and given the typical data collection proces
 s based on keywords\, available corpora are usually sparse and imbalanced 
 which makes the detection process challenging for current machine learning
  techniques.\n\nIn this talk\, I will present my findings when working on 
 (1) the construction of multilingual resources for robust toxic language a
 nd hate speech detection\, (2) the study of bias in toxic language detecti
 on\, and (3) the assessment of toxicity and harmful biases within Large Pr
 e-trained Language Models (PTLMs) which are at the core of major NLP syste
 ms.
LOCATION:Virtual (Zoom)
END:VEVENT
END:VCALENDAR