Variance Matters: Improving Domain Adaptation via Stratified Sampling

TL;DR

This paper introduces VaRDASS, a novel variance reduction technique for unsupervised domain adaptation that improves discrepancy estimation and target domain performance through stratified sampling.

Contribution

It develops the first specialized stochastic variance reduction method for UDA, with theoretical guarantees and a practical optimization algorithm.

Findings

Improved discrepancy estimation accuracy on four datasets.

Enhanced target domain performance with the proposed method.

Theoretical proof of variance minimization for MMD under certain conditions.

Abstract

Domain shift remains a key challenge in deploying machine learning models to the real world. Unsupervised domain adaptation (UDA) aims to address this by minimising domain discrepancy during training, but the discrepancy estimates suffer from high variance in stochastic settings, which can stifle the theoretical benefits of the method. This paper proposes Variance-Reduced Domain Adaptation via Stratified Sampling (VaRDASS), the first specialised stochastic variance reduction technique for UDA. We consider two specific discrepancy measures -- correlation alignment and the maximum mean discrepancy (MMD) -- and derive ad hoc stratification objectives for these terms. We then present expected and worst-case error bounds, and prove that our proposed objective for the MMD is theoretically optimal (i.e., minimises the variance) under certain assumptions. Finally, a practical k-means style optimisation algorithm is introduced and analysed. Experiments on four domain shift datasets demonstrate improved discrepancy estimation accuracy and target domain performance.