FairPOT: Balancing AUC Performance and Fairness with Proportional Optimal Transport

2025-08-05

AI Summary

In high-stakes domains (e.g., healthcare, finance, justice), achieving both fairness and high AUC performance remains challenging. Method: This paper proposes a tunable, model-agnostic post-processing framework based on *proportional optimal transport*, which selectively transforms risk scores via threshold-controllable calibration—specifically adjusting tail scores for disadvantaged groups—to improve inter-group AUC fairness while preserving global AUC nearly unchanged. Crucially, it extends optimization to partial AUC (pAUC), prioritizing fairness in high-risk score regions. Results: Experiments across multiple synthetic, public, and clinical datasets demonstrate substantial fairness gains with negligible—or even zero—AUC degradation. The method is computationally efficient, requires no model retraining, and is readily deployable in practice.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Stochastic Optimization
• Computer Vision: Bias, Fairness & Privacy

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

Fairness metrics utilizing the area under the receiver operator characteristic curve (AUC) have gained increasing attention in high-stakes domains such as healthcare, finance, and criminal justice. In these domains, fairness is often evaluated over risk scores rather than binary outcomes, and a common challenge is that enforcing strict fairness can significantly degrade AUC performance. To address this challenge, we propose Fair Proportional Optimal Transport (FairPOT), a novel, model-agnostic post-processing framework that strategically aligns risk score distributions across different groups using optimal transport, but does so selectively by transforming a controllable proportion, i.e., the top-lambda quantile, of scores within the disadvantaged group. By varying lambda, our method allows for a tunable trade-off between reducing AUC disparities and maintaining overall AUC performance. Furthermore, we extend FairPOT to the partial AUC setting, enabling fairness interventions to concentrate on the highest-risk regions. Extensive experiments on synthetic, public, and clinical datasets show that FairPOT consistently outperforms existing post-processing techniques in both global and partial AUC scenarios, often achieving improved fairness with slight AUC degradation or even positive gains in utility. The computational efficiency and practical adaptability of FairPOT make it a promising solution for real-world deployment.

Problem

• Balancing AUC performance and fairness in risk score models
• Reducing AUC disparities while maintaining overall AUC performance
• Extending fairness interventions to highest-risk regions with partial AUC

Innovation

• Uses optimal transport for fair risk score alignment
• Adjusts top quantile scores to balance fairness
• Extends to partial AUC for high-risk focus