Robust Importance-Weighted Cross-Validation under Sample Selection Bias

Abstract

Cross-validation under sample selection bias can, in principle, be done by importance-weighting the empirical risk. However, the importance-weighted risk estimator produces suboptimal hyperparameter estimates in problem settings where large weights arise with high probability. We study its sampling variance as a function of the training data distribution and introduce a control variate to increase its robustness to problematically large weights.