BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/1139
DTSTAMP:20230914T125952Z
SUMMARY:Leveraging the Invariance Principle for Out-of-Distribution General
 ization
DESCRIPTION:Speaker: Karthikeyan Shanmugan (IBM Research AI\nT.J. Watson Ce
 nter\, NY.)\n\nAbstract: \nOne of the fundamental issues facing deployment
  of supervised learning models in real life applications is the issue of o
 ut-of-distribution (OOD) generalization. Models trained using the standard
  Empirical Risk Minimization (ERM) on multiple training data sources suffe
 r from fitting to spurious features that correlate with label which does n
 ot hold in unseen test environments. ERM’s sole focus on optimizing aver
 age risk contributes to this problem. Invariance Principle\, in Pearlian C
 ausal Models\, has long been used to infer causal relationships from inter
 ventional data.\nInvariant Risk Minimization (IRM) is a recent paradigm th
 at proposes to leverage the invariance principle in an optimization framew
 ork for OOD problems. This paradigm views different training distributions
  and the unseen test as intervened versions of a common but unknown causal
  model. IRM seeks to identify that transformation of data such that the cl
 assifier trained on top of it is invariant across training domains apart f
 rom optimizing risk. Due to a challenging bilevel optimization\, a previou
 s proposal was limited to handling linear classifiers. We propose a novel 
 game theoretic learning paradigm – called Ensemble Invariant Risk Minimi
 zation (EIRM Game) whose Nash Equilibria is provably equivalent to invaria
 nt solutions for a very general class of non-linear classifiers and transf
 ormations. For least squares regression under unobserved confounding\, wit
 h a modified game we provide the first convergence guarantees\, known for 
 this problem in any setting\, to approximate invariant solutions (this par
 t may be discussed if time permits).\n\nBio: Karthikeyan Shanmugam is a Re
 search Staff Member with the IBM Research AI group in NY in the Trustworth
 y AI Department since 2017. Previously\, he was a Herman Goldstine Postdoc
 toral Fellow in the Mathematical Sciences Division at IBM Research\, NY. H
 e obtained his Ph.D. in Electrical and Computer Engineering from UT Austin
  in 2016\, MS degree in Electrical Engineering from USC in 2012 and B.Tech
 \, M.Tech degrees in Electrical Engineering from IIT Madras in 2010.\nHis 
 research interests broadly lie in Statistical Machine Learning (ML)\, Opti
 mization\, Graph Algorithms\, and Information Theory. In ML\, his focus is
  on causal inference\, online learning\, transfer learning and explainable
  ML. He has won several awards in IBM for his contributions to explainable
  AI and Causal Inference including the Corporate Technical Award in 2021\,
  the highest technical award in IBM. His works have appeared regularly in 
 top AI/ML venues like NeurIPS\, ICML\, AISTATS and ICLR.\n
URL:https://www.tcs.tifr.res.in/web/events/1139
DTSTART;TZID=Asia/Kolkata:20210628T090000
DTEND;TZID=Asia/Kolkata:20210628T100000
END:VEVENT
END:VCALENDAR