BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Causal Inference and Domain Adaptation - Jonas Peters\, ETH Zurich
DTSTART:20150129T130000Z
DTEND:20150129T140000Z
UID:TALK57305@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Why are we interested in the causal structure of a data-genera
 ting process? In a classical regression problem\, for example\, we include
  a variable into the model if it improves the prediction\; it seems that n
 o causal knowledge is required. In many situations\, however\, we are inte
 rested in the system's behaviour under a change of environment. Here\, cau
 sal models become important because they are usually considered invariant 
 under those changes. A causal prediction (which uses only direct causes of
  the target variable as predictors) remains valid even if we intervene on 
 predictor variables or change the whole experimental setting.\nWe introduc
 e structural equation models as a way of formalizing the invariance princi
 ple described above and present two ideas that can be used to infer causal
  structure from data: (1) restricted structural equation models and (2) a 
 recent method that exploits the invariance principle when data from differ
 ent environments are available.\nThis talk is meant as a short tutorial. I
 t concentrates on ideas and concepts and does not require any prior knowle
 dge about causality.\n
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station Road\, Cambridge
 \, CB1 2FB
END:VEVENT
END:VCALENDAR