BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Causal inference in high-dimensional systems based on observationa
 l data -  Marloes Maathuis  (ETH Zuerich)
DTSTART:20110211T160000Z
DTEND:20110211T170000Z
UID:TALK28562@talks.cam.ac.uk
CONTACT:Richard Nickl
DESCRIPTION:\nCause-effect relationships are of primary interest in many f
 ields of\nscience. We consider the problem of estimating such relationship
 s from\nobservational data\, that is\, from data obtained by observing the
  system\nof interest without subjecting it to interventions or perturbatio
 ns. We\ndiscuss that\, under some assumptions\, it is possible to consiste
 ntly\nestimate bounds on causal effects from such data\, even in high-dime
 nsional\nsettings where the number of variables is much larger than the sa
 mple\nsize. We present an experimental validation of our method\, using a\
 nchallenging high-dimensional yeast gene expression data set\, where\nwe c
 ould indeed find the strongest causal effects between genes.\nAn important
  application of this new type of statistical inference is that\nit offers 
 useful strategies for the prioritization of experiments.\n\n\nhttp://stat.
 ethz.ch/~maathuis/
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR