Cost Efficient Fairness Audit Under Partial Feedback

2025-10-04

AI Summary

This work addresses the efficient fairness auditing of classifiers under partial feedback—where only positively classified individuals receive ground-truth labels—a setting reflecting high labeling costs in real-world applications. To tackle this challenge, we propose two novel audit frameworks: one for black-box models and another for hybrid (partial-access) settings. Our methods integrate truncated sample learning, maximum a posteriori inference via an oracle, and exponential-family distribution modeling to estimate fairness metrics—including demographic parity, equal opportunity, and equalized odds—without full label access. Theoretically, we extend convergence analysis to exponential-family mixture distributions under partial feedback. Empirically, on benchmark datasets (Adult Income and Law School), our algorithms reduce auditing cost by approximately 50% on average compared to random sampling and state-of-the-art baselines, while maintaining statistical reliability.

Technology Category

• Machine Learning: Ethics, Bias, and Fairness
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Computer Vision: Bias, Fairness & Privacy

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Fairness and ethical considerations in crowd work and in human-in-the-loop AI systems
• Social Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

We study the problem of auditing the fairness of a given classifier under partial feedback, where true labels are available only for positively classified individuals, (e.g., loan repayment outcomes are observed only for approved applicants). We introduce a novel cost model for acquiring additional labeled data, designed to more accurately reflect real-world costs such as credit assessment, loan processing, and potential defaults. Our goal is to find optimal fairness audit algorithms that are more cost-effective than random exploration and natural baselines. In our work, we consider two audit settings: a black-box model with no assumptions on the data distribution, and a mixture model, where features and true labels follow a mixture of exponential family distributions. In the black-box setting, we propose a near-optimal auditing algorithm under mild assumptions and show that a natural baseline can be strictly suboptimal. In the mixture model setting, we design a novel algorithm that achieves significantly lower audit cost than the black-box case. Our approach leverages prior work on learning from truncated samples and maximum-a-posteriori oracles, and extends known results on spherical Gaussian mixtures to handle exponential family mixtures, which may be of independent interest. Moreover, our algorithms apply to popular fairness metrics including demographic parity, equal opportunity, and equalized odds. Empirically, we demonstrate strong performance of our algorithms on real-world fair classification datasets like Adult Income and Law School, consistently outperforming natural baselines by around 50% in terms of audit cost.

Problem

• Auditing classifier fairness with limited true label availability
• Developing cost-effective algorithms for partial feedback scenarios
• Optimizing fairness verification under real-world cost constraints

Innovation

• Novel cost model for partial feedback fairness audit
• Near-optimal black-box algorithm outperforming natural baselines
• Mixture model algorithm leveraging truncated samples and MAP