Robust Correction of Sampling Bias Using Cumulative Distribution Functions

TL;DR

This paper introduces a robust method for correcting covariate shift using empirical cumulative distribution functions, which is more stable and parameter-free compared to existing density ratio estimation techniques.

Contribution

The authors propose a new approach leveraging empirical CDFs for covariate shift correction, avoiding parameter tuning and improving robustness over current methods.

Findings

More robust predictions across datasets

No reliance on parameter tuning

Comparable classification performance to state-of-the-art

Abstract

Varying domains and biased datasets can lead to differences between the training and the target distributions, known as covariate shift. Current approaches for alleviating this often rely on estimating the ratio of training and target probability density functions. These techniques require parameter tuning and can be unstable across different datasets. We present a new method for handling covariate shift using the empirical cumulative distribution function estimates of the target distribution by a rigorous generalization of a recent idea proposed by Vapnik and Izmailov. Further, we show experimentally that our method is more robust in its predictions, is not reliant on parameter tuning and shows similar classification performance compared to the current state-of-the-art techniques on synthetic and real datasets.