BEGIN:VCALENDAR
VERSION:2.0
PRODID:talks.ox.ac.uk
BEGIN:VEVENT
SUMMARY:Uses of LLMs in Social Sciences: Benefits and Risk - Daniel Valden
 egro (University of Oxford)
DTSTART;VALUE=DATE-TIME:20260216T124500Z
DTEND;VALUE=DATE-TIME:20260216T140000Z
UID:https://talks.ox.ac.uk/talks/id/2d38d5db-5298-4f8c-8357-7c9d20d0a206/
DESCRIPTION:This presentation outlines a constrained use of large language
  models (LLMs) in sociology and demography\, grounded in an information-th
 eoretic account of language. LLMs are treated as lossy statistical compres
 sion systems operating over non-injective mappings from social meaning to 
 text.\n\nOn this basis\, their appropriate uses are limited to supporting 
 existing analytical work: improving clarity of expression\, surfacing text
 ual regularities\, and stress-testing arguments and assumptions. They are 
 not sources of evidence\, explanation\, or social inference. Used within t
 hese limits\, LLMs can reduce linguistic friction without being mistaken f
 or epistemic agents.\n\nBiography:\nDaniel is a Senior Data Scientist and 
 Postdoctoral Researcher in Computational Social Science at the Leverhulme 
 Centre for Demographic Science. His research in the Centre is focused on t
 he development of robust estimation methods for social science and in the 
 development of software libraries in Python and R to perform multiverse-ty
 pe estimations.\n\nAdditionally\, he researches the application of machine
  learning / deep learning models (e.g.\, BERT\, RoBERTa\, GPT-2) on social
  science problems like misinformation detection and characterisation on so
 cial media text\, and the characterisation of social movement emotions ove
 r time based on associated tweets. \n\nPrior to joining Oxford\, Daniel co
 mpleted his PhD in Computational Social Science at the University of Leeds
 \, and before that worked for several years as a quantitative analyst at t
 he Pontifical Catholic University of Chile. His general interests are rela
 ted to the use of machine learning methods to understand human behaviour a
 nd the application of novel methods for robust parameter estimation\, eith
 er using multiverse-type approaches or Bayesian / probabilistic approaches
 . \n\nPlease join either in person or online. For in-person attendees\, th
 e talk will be preceded by a light lunch at 12.15pm.\nSpeakers:\nDaniel Va
 ldenegro (University of Oxford)
LOCATION:Seminar Room\, Department of Sociology and Online
TZID:Europe/London
URL:https://talks.ox.ac.uk/talks/id/2d38d5db-5298-4f8c-8357-7c9d20d0a206/
BEGIN:VALARM
ACTION:display
DESCRIPTION:Talk:Uses of LLMs in Social Sciences: Benefits and Risk - Dani
 el Valdenegro (University of Oxford)
TRIGGER:-PT1H
END:VALARM
END:VEVENT
END:VCALENDAR