Fairness Interventions in Classification: A Study on AI Explainability

A new study from arXiv argues that Equalized Odds is a more reliable fairness criterion than Demographic Parity for guiding bias correction in AI classification. The researchers introduce FairDream, a fairness package that increases model weights on errors for disadvantaged groups, and benchmark it against the GridSearch method. The study provides a normative justification for Equalized Odds while acknowledging its limitations.

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.