Machine Learning with Multitype Protected Attributes: Intersectional Fairness through Regularisation

Paper: arXiv 2509.08163

Authors: Ho Ming Lee, Katrien Antonio, Benjamin Avanzi, Lorenzo Marchi, Rui Zhou

Abstract

Ensuring equitable treatment (fairness) across protected attributes (such as gender or ethnicity) is a critical issue in machine learning. Most existing literature focuses on binary classification, but achieving fairness in regression tasks-such as insurance pricing or hiring score assessments-is equally important. Moreover, anti-discrimination laws also apply to continuous attributes, such as age, for which many existing methods are not applicable. In practice, multiple protected attributes can exist simultaneously; however, methods targeting fairness across several attributes often overlook so-called “fairness gerrymandering”, thereby ignoring disparities among intersectional subgroups (e.g., African-American women or Hispanic men). In this paper, we propose a distance covariance regularisation framework that mitigates the association between model predictions and protected attributes, in line with the fairness definition of demographic parity, and that captures both linear and nonlinear dependencies. To enhance applicability in the presence of multiple protected attributes, we extend our framework by incorporating two multivariate dependence measures based on distance covariance: the previously proposed joint distance covariance (JdCov) and our novel concatenated distance covariance (CCdCov), which effectively address fairness gerrymandering in both regression and classification tasks involving protected attributes of various types. We discuss and illustrate how to calibrate regularisation strength, including a method based on Jensen-Shannon divergence, which quantifies dissimilarities in prediction distributions across groups. We apply our framework to the COMPAS recidivism dataset and a large motor insurance claims dataset.

Complexity vs Empirical Score

• Math Complexity: 6.5/10
• Empirical Rigor: 7.5/10
• Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper provides a novel distance covariance regularization framework (incorporating JdCov and a novel CCdCov) to tackle intersectional fairness and gerrymandering across multiple continuous and discrete attributes. It combines strong statistical foundations with rigorous empirical applications on insurance claims and recidivism benchmark datasets.

Research Flowchart

  flowchart TD
    A[Goal: Intersectional Fairness in ML<br>Regression & Classification] --> B[Input Data: COMPAS Recidivism<br>& Motor Insurance Claims]
    B --> C[Method: Distance Covariance<br>Regularisation Framework]
    C --> D[Multivariate Dependence Extensions:<br>JdCov & Novel CCdCov]
    D --> E[Calibrate Regularisation Strength<br>via Jensen-Shannon Divergence]
    E --> F[Outcomes: Mitigated Linear/Nonlinear Disparities<br>& Prevented Fairness Gerrymandering]

Cited References (110)

1. Fairness in Clustering with Multiple Sensitive Attributes (2020) — link
2. Machine bias risk assessments in criminal sentencing (2016) — link
3. Guidance Resource: Artificial Intelligence and Discrimination in Insurance Pricing and Underwriting (2022) — link
4. Understanding Insurance Antidiscrimination Laws (2014) — link
5. Fairness and Machine Learning: Limitations and Opportunities (2023) — link
6. Penalizing Unfairness in Binary Classification (2017) — link
7. A Convex Framework for Fair Regression (2017) — link
8. A Survey of Bias in Machine Learning through the Prism of Statistical Parity (2022) — link

0 of 110 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.