BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:NLIP Seminar Series
SUMMARY:Understanding the Interplay between LLMs' Utilisat
 ion of Parametric and Contextual Knowledge  - Prof
  Isabelle Augenstein (University of Copenhagen)
DTSTART;TZID=Europe/London:20260501T120000
DTEND;TZID=Europe/London:20260501T130000
UID:TALK244069AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/244069
DESCRIPTION:Language Models (LMs) acquire parametric knowledge
  from their training process\, embedding it within
  their weights. The increasing scalability of LMs\
 , however\, poses significant challenges for under
 standing a model's inner workings and further for 
 updating or correcting this embedded knowledge wit
 hout the significant cost of retraining. Moreover\
 , when using these language models for knowledge-i
 ntensive language understanding tasks\, LMs have t
 o integrate relevant context\, mitigating their in
 herent weaknesses\, such as incomplete or outdated
  knowledge. Nevertheless\, studies indicate that L
 Ms often ignore the provided context as it can be 
 in conflict with the pre-existing LM's memory lear
 ned during pre-training. Conflicting knowledge can
  also already be present in the LM's parameters\, 
 termed intra-memory conflict. This underscores the
  importance of understanding the interplay between
  how a language model uses its parametric knowledg
 e and the retrieved contextual knowledge.\n\nIn th
 is talk\, I will aim to shed light on this importa
 nt issue by presenting our research on evaluating 
 the knowledge present in LMs\, diagnostic tests th
 at can reveal knowledge conflicts\, as well as on 
 understanding the characteristics of successfully 
 used contextual knowledge.\n\n\n\n**Bio:** Isabell
 e Augenstein is a Professor at the University of C
 openhagen\, Department of Computer Science\, where
  she heads the Natural Language Processing section
 . Her main research interests are fair and account
 able NLP\, including challenges such as explainabi
 lity\, factuality and bias detection. Prior to sta
 rting a faculty position\, she was a postdoctoral 
 researcher at University College London\, and befo
 re that a PhD student at the University of Sheffie
 ld. In October 2022\, Isabelle Augenstein became D
 enmark’s youngest ever female full professor. She 
 currently holds a prestigious ERC Starting Grant o
 n 'Explainable and Robust Automatic Fact Checking’
 \, and her research has been recognised by a Karen
  Spärck Jones Award\, as well as a Hartmann Diplom
 a Prize. She is a member of the Royal Danish Acade
 my of Sciences and Letters\, and co-leads the Dani
 sh Pioneer Centre for AI.
LOCATION:FW11 Hybrid (In-Person + Online). Here is the Goog
 le Meet Link: https://meet.google.com/cru-hcuo-rhu
CONTACT:Suchir Salhan
END:VEVENT
END:VCALENDAR