aiqbench analysis
This research evaluates fairness correction tools for AI classification models, arguing that Equalized Odds is a more robust criterion than Demographic Parity for guiding bias mitigation. The study introduces FairDream, an accessible package designed to reweight errors on disadvantaged groups while maintaining transparency in its corrective mechanisms. By benchmarking against stricter parity enforcement methods, the authors provide normative justification for conditioning counterfactual evaluations on true labels to address Simpson's paradox-like structures in fairness metrics.
Fairness Interventions in Classification: A Study on AI Explainability
arXiv:2407.14766v4 Announce Type: replace-cross Abstract: This paper presents a philosophical and experimental study of fairness interventions in AI classification, centered on the explainability and transparency of corrective methods, and on the opposition between two fairness criteria, namely Demographic Parity and Equalized Odds. Our main argument is that even as a gap in Demographic Parity is used to diagnose inequality between groups, Equalized Odds constitutes a more reliable fairness criterion to guide bias correction in classification. To establish this, we present FairDream, a fairness package intended for lay users, whose mechanism increases the model's weights of errors on disadvantaged groups. To justify FairDream's results, we analyze its reweighting algorithm, and we present the results of a benchmark experiment in which we compare FairDream with a distinct in-processing correction method that enforces Demographic Parity more drastically, the GridSearch method. We then propose a normative justification of Equalized Odds, with a discussion of the criterion's limitations. We draw on the structural similarity between FairDream's results and a version of Simpson's paradox to justify conditioning on true labels in counterfactual evaluations of fairness.
Abstract:This paper presents a philosophical and experimental study of fairness interventions in AI classification, centered on the explainability and transparency of corrective methods, and on the opposition between two fairness criteria, namely Demographic Parity and Equalized Odds. Our main argument is that even as a gap in Demographic Parity is used to diagnose inequality between groups, Equalized Odds constitutes a more reliable fairness criterion to guide bias correction in classification. To establish this, we present FairDream, a fairness package intended for lay users, whose mechanism increases the model's weights of errors on disadvantaged groups. To justify FairDream's results, we analyze its reweighting algorithm, and we present the results of a benchmark experiment in which we compare FairDream with a distinct in-processing correction method that enforces Demographic Parity more drastically, the GridSearch method. We then propose a normative justification of Equalized Odds, with a discussion of the criterion's limitations. We draw on the structural similarity between FairDream's results and a version of Simpson's paradox to justify conditioning on true labels in counterfactual evaluations of fairness.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
I.2.6; K.4.1; K.4.2
arXiv:2407.14766 [cs.LG]
(or arXiv:2407.14766v4 [cs.LG] for this version)
Full text under the publisher's open licence · 195 words
Related coverage
SCM-based Fairness and Faithful Explainability for Legal Document Classification
arXiv · Oct 2, 2026
The Entangled Dimensions of Decision Theory
LessWrong · Jul 30, 2026
Integrating Fairness and Explainability in a Multiple Instance Reinforcement Learning System
arXiv · Oct 2, 2026
Unpacking AI safety for enterprises
Cohere · Jun 4, 2026