BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Deconfounding using Spectral Transformations - Domagoj Ćevid (ETH
  Zürich)\; Peter Bühlmann (ETH Zürich)
DTSTART:20180629T104500Z
DTEND:20180629T113000Z
UID:TALK107527@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:High-dimensional regression methods which rely on the sparsity
  of the ground truth\, such as the Lasso\, might break down in the presenc
 e of confounding variables. If a latent variable affects both the response
  and the predictors\, the correlation between them changes. This phenomeno
 n can be represented as a linear model where the sparse coefficient vector
  has been perturbed. We will present our work on this problem. We investig
 ate and propose some spectral transformations for the data which serve as 
 input for the Lasso. We discuss assumptions for achieving the optimal erro
 r rate and illustrate the performance on a genomic dataset. The approach i
 s easy to use and leads to convincing results. The talk is based on joint 
 work with Nicolai Meinshausen.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR