Avoiding Discrimination through Causal Reasoning

TL;DR

This paper advocates for causal reasoning in fairness assessment in machine learning, highlighting limitations of observational criteria and proposing causal non-discrimination criteria with algorithms to enforce them.

Contribution

It introduces a causal framework for fairness, formalizes the limitations of observational criteria, and develops algorithms for causal non-discrimination.

Findings

Observational criteria often fail to ensure fairness.

Causal reasoning clarifies when and why discrimination occurs.

Algorithms for causal non-discrimination are proposed.

Abstract

Recent work on fairness in machine learning has focused on various statistical discrimination criteria and how they trade off. Most of these criteria are observational: They depend only on the joint distribution of predictor, protected attribute, features, and outcome. While convenient to work with, observational criteria have severe inherent limitations that prevent them from resolving matters of fairness conclusively. Going beyond observational criteria, we frame the problem of discrimination based on protected attributes in the language of causal reasoning. This viewpoint shifts attention from "What is the right fairness criterion?" to "What do we want to assume about the causal data generating process?" Through the lens of causality, we make several contributions. First, we crisply articulate why and when observational criteria fail, thus formalizing what was before a matter of opinion. Second, our approach exposes previously ignored subtleties and why they are fundamental to the problem. Finally, we put forward natural causal non-discrimination criteria and develop algorithms that satisfy them.