BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Extrapolation-aware statistical machine learning - Peter Bühlmann
  (ETH Zurich)
DTSTART:20250509T130000Z
DTEND:20250509T140000Z
UID:TALK231163@talks.cam.ac.uk
CONTACT:Qingyuan Zhao
DESCRIPTION:Nonparametric function estimation and prediction with moderate
  or large dimension of the covariates are particularly susceptible to extr
 apolation\, because data points are typically far apart from each other in
  such moderate or higher dimension. Thus\, there is a need to have machine
  learning methods that are extrapolation-aware\, i.e. that automatically p
 erform well (in a sense) when extrapolation occurs. Without such extrapola
 tion-aware techniques\, inference from standard machine learning and nonpa
 rametric procedures may be poor or invalid. We introduce a novel conceptua
 l framework and introduce *Xtrapolation* which allows for extrapolation-aw
 are inference with any ML algorithm. \n\nThis is joint work with Niklas Pf
 ister (Lakera AI)
LOCATION:MR12\, Centre for Mathematical Sciences
END:VEVENT
END:VCALENDAR