BEGIN:VCALENDAR
PRODID:-//Columba Systems Ltd//NONSGML CPNG/SpringViewer/ICal Output/3.3-
 M3//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20241028T115635Z
DTSTART:20241030T110000Z
DTEND:20241030T120000Z
SUMMARY:AI-Fun with ELLIS Seminar | Chrysoula Zerva: Uncertainty in NLP &
  beyond: quantification\, interpretation\, evaluation
UID:{http://www.columbasystems.com/customers/uom/gpp/eventid/}ai9-lzv6fdf
 w-ixt0s8
DESCRIPTION:The Manchester Centre for AI Fundamentals and Manchester's EL
 LIS Unit are co-hosting a series of seminars featuring expert researcher
 s working in the fundamentals of AI. \n\nTitle: Uncertainty in NLP & bey
 ond: quantification\, interpretation\, evaluation\n\nAbstract:\nAs langu
 age models grow in popularity\, size\, and application across a wide ran
 ge of tasks\, they are becoming ubiquitous in modern society. This\, in 
 turn\, brings forward questions on reliability\, trust and interpretabil
 ity. We know models don’t always “know what they don’t know” and may end
  up generating seemingly convincing answers that are entirely fabricated
 . Hence\, obtaining reliable confidence estimates\, i.e. being able to q
 uantify the uncertainty over their predictions\, is a key step in the pa
 th towards the reliability of language models and more responsible AI so
 lutions.\n\nThis talk will discuss the challenges of uncertainty estimat
 ion for natural language processing\, emphasising aspects such as multip
 le sources of uncertainty and limited access to the model parameters (bl
 ack box models) as well as aspects of interpretation and evaluation. I w
 ill focus on generation and evaluation tasks\, using machine translation
  as the main paradigm\, and discuss how the conformal prediction framewo
 rk can be leveraged to provide meaningful confidence intervals with stat
 istical guarantees\, while also allowing us to calibrate our confidence 
 to obtain more interpretable and fair uncertainty representations.\n\nCh
 rysoula (Chryssa) Zerva is an Assistant Professor for Artificial Intelli
 gence at IST and a researcher at Instituto de Telecomunicações (IT) in L
 isbon. She is a member of the European Laboratory for Learning & Intelli
 gent Systems (ELLIS) and LUMLIS\, the Lisbon Unit for Learning & Intelli
 gent Systems. She obtained her PhD in 2019 from the University of Manche
 ster on "Automated Identification of Textual Uncertainty". In 2019 she a
 lso received the EPSRC Doctoral Prize Fellowship\, and in 2021\, she joi
 ned IT as a postdoc for the DeepSPIN project.\nShe is a co-PI in the Cen
 tre for Responsible AI project\, a PRR initiative for trustworthy\, sust
 ainable\, fair\, and transparent AI. She is also part of the UTTER proje
 ct\, with her focus being on uncertainty-aware\, adaptable and context-a
 ware models. Her research focuses on elucidating uncertainty in machine 
 learning and especially NLP. She also explores topics on explainability\
 , fairness and quality estimation under multilingual and multimodal setu
 ps. \n\n
STATUS:TENTATIVE
TRANSP:TRANSPARENT
CLASS:PUBLIC
LOCATION:Lecture Theatre 1.5\, Kilburn Building
END:VEVENT
END:VCALENDAR