BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20261031T052807EDT-215994WKoe@132.216.98.100
DTSTAMP:20261031T092807Z
DESCRIPTION:Hengrui Cai\, PhD\n\nAssistant Professor of Statistics\n	Donald 
 Bren School of Information\n	and Computer Sciences | UCI\n\nWHEN: Wednesday
 \, January 8\, 2025\, from 3:30 to 4:30 p.m.\n	WHERE: Hybrid | 2001 McGill 
 College Avenue\, Room 1140\; Zoom\n	NOTE: Hengrui Cai will be presenting vi
 rtually\n\nAbstract\n\nThe causal revolution has spurred interest in under
 standing complex relationships across various fields. Most existing method
 s aim to discover causal relationships among variables within a complex\, 
 large-scale system. However\, in practice\, only a small number of variabl
 es are relevant to the outcomes of interest. As a result\, causal estimati
 on using the full causal representation\, especially with limited data\, c
 ould lead to many falsely discovered\, spurious variables that are highly 
 correlated with but have no causal impact on the target outcome. We propos
 e learning a class of necessary and sufficient causal graphs that only con
 tain causally relevant variables\, utilizing probabilities of causation. T
 he framework is further extended to natural language processing models to 
 disentangle the 'black box' by identifying true rationales when two or mor
 e snippets are highly inter-correlated\, thus contributing similarly to pr
 ediction accuracy. We leverage two causal desiderata\, non-spuriousness an
 d efficiency\, establishing their theoretical identification as the main c
 omponent of learning necessary and sufficient in language models. The supe
 rior performance of our proposed methods is demonstrated in real-world rev
 iews and medical datasets through extensive experiments.\n\n \n\nSpeaker B
 io\n\nDr. Hengrui Cai is an Assistant Professor in the Department of Stati
 stics in the DonaldBren School of Information and Computer Sciences at the
  University of CaliforniaIrvine. She obtained her Ph.D. degree in Statisti
 cs at North Carolina State University in2022. Cai has broad research inter
 ests in methodology and theory in causal inference\,reinforcement learning
 \, and graphical modeling\, to establish reliable\, powerful\, andinterpre
 table solutions to real-world problems. Currently\, her research focuses o
 ncausal inference and causal structure learning\, natural language process
 ing andexplainable deep learning\, and policy optimization and evaluation 
 in reinforcementlearning and bandits\, with applications in precision medi
 cine. Her work has beenpublished in journals including the Journal of the 
 American Statistical Association\,Journal of Machine Learning Research\, a
 nd Statistics in Medicine\, as well asconferences including NeurIPS\, ICML
 \, and ICLR. Please visit her personal website: https://hengruicai.github.
 io/.\n\n \n
DTSTART:20250108T203000Z
DTEND:20250108T213000Z
SUMMARY:Towards Trustworthy Machine Learning: A Causal Lens on Learning Non
 -Spuriousness
URL:https://www.mcgill.ca/epi-biostat-occh/channels/event/towards-trustwort
 hy-machine-learning-causal-lens-learning-non-spuriousness-362129
END:VEVENT
END:VCALENDAR