Explainability for fair machine learning

TL;DR

This paper introduces a Shapley value-based explainability method for assessing and understanding fairness in machine learning models, even when sensitive attributes are not directly used, and proposes a meta algorithm for fairness interventions.

Contribution

It presents a novel Shapley value approach to explain model fairness and a meta algorithm for training fairness interventions without sacrificing performance.

Findings

The method effectively attributes unfairness to input features.

The meta algorithm improves fairness with no performance loss.

Explains trade-offs between accuracy and fairness.

Abstract

As the decisions made or influenced by machine learning models increasingly impact our lives, it is crucial to detect, understand, and mitigate unfairness. But even simply determining what "unfairness" should mean in a given context is non-trivial: there are many competing definitions, and choosing between them often requires a deep understanding of the underlying task. It is thus tempting to use model explainability to gain insights into model fairness, however existing explainability tools do not reliably indicate whether a model is indeed fair. In this work we present a new approach to explaining fairness in machine learning, based on the Shapley value paradigm. Our fairness explanations attribute a model's overall unfairness to individual input features, even in cases where the model does not operate on sensitive attributes directly. Moreover, motivated by the linearity of Shapley explainability, we propose a meta algorithm for applying existing training-time fairness interventions, wherein one trains a perturbation to the original model, rather than a new model entirely. By explaining the original model, the perturbation, and the fair-corrected model, we gain insight into the accuracy-fairness trade-off that is being made by the intervention. We further show that this meta algorithm enjoys both flexibility and stability benefits with no loss in performance.