Learning Fair Domain Adaptation with Virtual Label Distribution

TL;DR

This paper introduces VILL, a framework for unsupervised domain adaptation that improves category fairness by adaptively re-weighting hard-to-classify categories and explicitly balancing decision boundaries, leading to more equitable performance across categories.

Contribution

VILL is a novel, plug-and-play framework that enhances category fairness in UDA by combining adaptive re-weighting and KL-divergence-based re-balancing strategies.

Findings

VILL significantly improves worst-case category performance.

VILL maintains high overall accuracy while enhancing fairness.

VILL can be integrated into existing UDA methods with ease.

Abstract

Unsupervised Domain Adaptation (UDA) aims to mitigate performance degradation when training and testing data are sampled from different distributions. While significant progress has been made in enhancing overall accuracy, most existing methods overlook performance disparities across categories-an issue we refer to as category fairness. Our empirical analysis reveals that UDA classifiers tend to favor certain easy categories while neglecting difficult ones. To address this, we propose Virtual Label-distribution-aware Learning (VILL), a simple yet effective framework designed to improve worst-case performance while preserving high overall accuracy. The core of VILL is an adaptive re-weighting strategy that amplifies the influence of hard-to-classify categories. Furthermore, we introduce a KL-divergence-based re-balancing strategy, which explicitly adjusts decision boundaries to enhance category fairness. Experiments on commonly used datasets demonstrate that VILL can be seamlessly integrated as a plug-and-play module into existing UDA methods, significantly improving category fairness.