equity impact analysis

This study addresses the critical gap in clinical machine learning fairness evaluation by systematically applying an intersectional fairness auditing framework to real-world clinical prediction tasks. Leveraging the All of Us dataset, the authors integrate the FairLogue toolkit, observational fairness metrics, and counterfactual causal analysis to assess model performance across intersecting subgroups defined by race and gender. Their findings reveal substantial performance disparities that remain undetected under conventional single-axis fairness assessments. However, counterfactual experiments demonstrate that most of these disparities persist even after randomizing group identity, indicating that they primarily stem from differences in covariate distributions rather than direct discrimination. These results underscore the necessity and value of intersectional auditing for accurately diagnosing and addressing health inequities in clinical AI systems.

Machine learning models often amplify inherent societal biases in data, disproportionately harming minority groups. This paper addresses fairness risks at the earliest stage of the ML pipeline—data collection—and introduces the first fairness risk auditing framework grounded in Algorithmic Information Theory (AIT). Our method quantifies bias risk via AIT, integrating statistical bias metrics with cross-dataset benchmarking to jointly achieve three objectives: (i) data-level risk prediction, (ii) identification of model families sensitive to bias, and (iii) interpretable attribution of risk sources. Experiments on standard fairness benchmarks demonstrate that the framework accurately identifies high-risk datasets and bias-prone model families, significantly improving early detection rates and interpretability of bias risks. By enabling proactive, data-centric fairness assessment, it provides a deployable, pre-modeling audit tool for fair ML development.

This work addresses a critical gap in algorithmic fairness by observing that existing fairness metrics predominantly focus on disparities in model outcomes across groups while overlooking biases inherent in the explanation process itself. To bridge this gap, the authors propose Group-wise Explanation Stability Discrepancy (GESD), an explainer-agnostic and model-agnostic, procedure-oriented fairness metric. They further introduce the Fairness–Explanation–Utility (FEU) framework, which jointly optimizes predictive utility, outcome fairness, and explanation fairness through multi-objective optimization and explainability techniques. This approach is the first to extend fairness evaluation to the level of explanations, enabling comparative analysis of explanation stability across demographic groups. Empirical results demonstrate that GESD effectively captures group-based disparities in explanation quality, and the FEU framework consistently achieves superior trade-offs between prediction accuracy and fairness across multiple benchmark datasets compared to state-of-the-art methods.

Existing fair machine learning methods overlook the impact of human decision-making on downstream outcomes, thus failing to mitigate real-world outcome inequality arising from representational bias. To address this, we introduce the novel concept of “representation mismatch,” formally modeling the causal discrepancy between actual human decisions and idealized counterfactual decisions. We cast this mismatch as an intervenable multi-objective optimization problem within a neural network framework. Theoretically, our learned interpretable weights provably eliminate downstream outcome inequality, enabling decision-maker–oriented explainable interventions (e.g., behavioral nudges). Our method integrates causal simplifying assumptions with weight-based representation modeling. Empirical evaluation on German Credit, Adult, and Heritage Health datasets confirms both the identifiability and remediability of representation mismatch, achieving complete mitigation of downstream inequality.

This paper addresses the misalignment between algorithmic bias assessment and legal standards by proposing a quantification framework rigorously grounded in U.S. anti-discrimination law. Methodologically, it distinguishes legally salient discriminatory testing from systemic disparity through legal contextualization, and introduces the Objective Fairness Index (OFI)—a metric integrating objective test theory and measurement stability, using marginal benefit as a proxy to quantify legal compliance of algorithmic decisions. Its key contribution lies in being the first fairness metric to embed legal admissibility directly into its design, enabling a paradigm shift in algorithmic auditing from statistical fairness to legally grounded fairness. Empirical evaluation on real-world judicial prediction systems—including COMPAS—demonstrates that OFI reliably detects unlawful discrimination, offering regulators and auditors the first quantitative tool with both legal interpretability and operational utility.

This study addresses how infra-marginality—differences in data distributions across groups—complicates judgments of AI fairness, as conventional statistical parity metrics often fail to align with human perceptions of fairness. Through a controlled user study involving 85 participants in a hypothetical medical decision-making scenario, the authors systematically investigate how group-specific model performance and training data availability shape fairness judgments. They find that when group-wise performance is equal or unknown, participants favor outcome equality; however, when performance disparities are attributable to data imbalance, models preserving these differences are perceived as more fair. These results demonstrate that human fairness judgments are not solely based on outcome equality but are significantly influenced by beliefs about the underlying causes of disparities, thereby challenging the prevailing assumption that statistical parity should serve as the default standard for algorithmic fairness.

This work addresses the challenge of identifying bias sources under domain shift, where conventional scalar fairness metrics often fall short. The authors propose RISE, a novel diagnostic framework that, for the first time, formally links ranking residual curves with rigorous fairness notions to enable fine-grained and interactive fairness analysis. By leveraging residual ranking, visualization, and post-hoc examination, RISE facilitates localized fairness evaluation, cross-environment subgroup comparisons, and the discovery of hidden biases. The framework effectively uncovers accuracy–fairness trade-offs obscured by aggregate statistics, thereby empowering practitioners to make more informed and equitable model selection decisions.

Algorithmic decision-making often faces a trade-off between fairness and interpretability. Method: This paper proposes a synergistic optimization framework that integrates sensitive-attribute decorrelation preprocessing with interpretable policy trees. It introduces a novel feature-space inverse transformation mechanism that mitigates the influence of sensitive attributes on decisions while preserving original feature semantics—ensuring policy transparency—and enhances fairness and prediction stability through structural tree optimization. Contribution/Results: Evaluated on Swiss labor market policy allocation, the method significantly improves group-level fairness—e.g., statistical parity increases by 23%—while incurring only a marginal reduction in employment rate (<1.5%). These results demonstrate its effectiveness and practical viability in real-world policy deployment.

This work addresses the limitations of existing fairness methods, which often focus on a single demographic attribute and lack systematic evaluation across intersecting subgroups and multiple stages of the modeling pipeline. To bridge this gap, we propose FairSelect, a novel toolkit that establishes the first multi-level evaluation framework enabling arbitrary combinations of pre-, in-, and post-processing fairness interventions. We conduct comprehensive analyses of fairness–utility trade-offs across diverse model architectures and intersectional subgroups using both synthetic clinical data and a real-world atrial fibrillation stroke risk prediction task. Our experiments demonstrate that combined intervention strategies generally enhance fairness with controllable utility loss; notably, certain combinations simultaneously improve both fairness and predictive performance, while others yield adverse effects, revealing non-additive and context-dependent interactions among fairness interventions in intersectional settings.

Existing fairness tools are often limited to single demographic attributes and struggle to capture the compounded biases faced by intersecting groups—such as combinations of race and gender—in clinical machine learning. This work proposes a Python toolkit that extends observational fairness metrics, including demographic parity and equalized odds, to intersectional subgroups for the first time, while integrating two counterfactual fairness frameworks to evaluate intervention-based equity. Applied to electronic health record data using logistic regression in a glaucoma surgery prediction task, the approach uncovers substantial intersectional unfairness, with a demographic parity gap as high as 0.20. Crucially, disparities identified through intersectional analysis markedly exceed those detected by single-dimension assessments, underscoring the necessity and efficacy of this method for auditing fairness in clinical algorithms.