Extreme Dimension Reduction for Handling Covariate Shift
This addresses covariate shift in machine learning, but the results are incremental as the method only works sometimes.
The paper tackled the problem of covariate shift by exploring extreme dimension reduction to improve importance weighting methods, showing that it sometimes helps but not always due to introduced bias.
In the covariate shift learning scenario, the training and test covariate distributions differ, so that a predictor's average loss over the training and test distributions also differ. In this work, we explore the potential of extreme dimension reduction, i.e. to very low dimensions, in improving the performance of importance weighting methods for handling covariate shift, which fail in high dimensions due to potentially high train/test covariate divergence and the inability to accurately estimate the requisite density ratios. We first formulate and solve a problem optimizing over linear subspaces a combination of their predictive utility and train/test divergence within. Applying it to simulated and real data, we show extreme dimension reduction helps sometimes but not always, due to a bias introduced by dimension reduction.