BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Numerical Reasoning in Natural Language Processing - Nafise Moosav
 i (University of Sheffield)
DTSTART:20231110T120000Z
DTEND:20231110T130000Z
UID:TALK206200@talks.cam.ac.uk
CONTACT:Michael Schlichtkrull
DESCRIPTION:Numerical reasoning is a fundamental skill for language models
  to understand textual input in tasks such as text generation\, question a
 nswering\, and fact checking. The predominant approach for enhancing numer
 ical reasoning has been scaling—larger models trained on more data tend 
 to perform better on relevant benchmarks that require numerical reasoning.
  Unfortunately\, this solution is often not accessible to the majority of 
 end users of such models\, as it is typically available through paid APIs 
 or prominent research labs. In this presentation\, we address the challeng
 es of end-to-end numerical reasoning in various natural language processin
 g (NLP) tasks. These challenges encompass evaluating the underlying reason
 ing skills while performing downstream applications and finding ways to im
 prove these skills without resorting to scaling.\n\nDr. Nafise Sadat Moosa
 vi is a Lecturer in Natural Language Processing at the Computer Science De
 partment of the University of Sheffield. Before joining the University of 
 Sheffield\, she was a postdoctoral researcher at the Technical University 
 of Darmstadt. She works on the limitation of language models to improve th
 eir fairness\, reasoning\, robustness\, and efficiency. She co-founded and
  co-organizes SustaiNLP workshops and regularly serves as a senior area ch
 air and area chair at *ACL conferences.\n\n\n
LOCATION:Computer Lab\, SS03
END:VEVENT
END:VCALENDAR