Robust Learning via Cause-Effect Models

Bernhard Schölkopf, Dominik Janzing, Jonas Peters +1

We consider the problem of function estimation in the case where the data distribution may shift between training and test time, and additional information about it may be availabl…