Cross-model Fairness: Empirical Study of Fairness and Ethics Under Model Multiplicity
Kacper Sokol, Meelis Kull, Jeffrey Chan, Flora Salim
TL;DR
This work introduces cross-model fairness as fairness under model multiplicity, formalizing a collection $\mathcal{F}_\epsilon$ of performance-equivalent predictors and disputable spaces where predictions diverge. It combines theoretical definitions with an extensive empirical study on OpenML pipelines across Credit Approval, German Credit, and Adult datasets, using ambiguity and discrepancy metrics and novel visualization tools to quantify unfairness arising from model multiplicity. The authors demonstrate that enforcing a fair-by-design ensemble—always selecting the most favourable outcome from $\mathcal{F}_\epsilon$—can degrade overall predictive performance, especially for expressive models, and that restricting the model space or choosing alternative aggregation schemes can help mitigate unfairness. Overall, the paper highlights a new dimension of fairness relevant to real-world ML systems, motivates practical safeguards (model-space constraints, abstention, human-in-the-loop), and opens avenues for rigorous bounds and alternative aggregation methods to handle disputable regions.
Abstract
While data-driven predictive models are a strictly technological construct, they may operate within a social context in which benign engineering choices entail implicit, indirect and unexpected real-life consequences. Fairness of such systems -- pertaining both to individuals and groups -- is one relevant consideration in this space; algorithms can discriminate people across various protected characteristics regardless of whether these properties are included in the data or discernible through proxy variables. To date, this notion has predominantly been studied for a fixed model, often under different classification thresholds, striving to identify and eradicate undesirable, discriminative and possibly unlawful aspects of its operation. Here, we backtrack on this fixed model assumption to propose and explore a novel definition of cross-model fairness where individuals can be harmed when one predictor is chosen ad hoc from a group of equally well performing models, i.e., in view of utility-based model multiplicity. Since a person may be classified differently across models that are otherwise considered equivalent, this individual could argue for a predictor granting them the most favourable outcome, employing which may have adverse effects on other people. We introduce this scenario with a two-dimensional example and linear classification; then, we present a comprehensive empirical study based on real-life predictive models and data sets that are popular with the algorithmic fairness community; finally, we investigate analytical properties of cross-model fairness and its ramifications in a broader context. Our findings suggest that such unfairness can be readily found in real life and it may be difficult to mitigate by technical means alone as doing so is likely to degrade predictive performance.