FairOD: Fairness-aware Outlier Detection

TL;DR

FairOD introduces a novel fairness-aware outlier detection method that ensures treatment parity and group fairness while maintaining high detection accuracy, addressing a significant gap in fair machine learning for unsupervised outlier detection.

Contribution

This work formalizes fairness criteria for outlier detection and proposes FairOD, the first method to incorporate fairness constraints into OD, achieving fair and effective detection.

Findings

FairOD achieves treatment parity and group fairness in outlier detection.

FairOD performs comparably or better than fairness-agnostic detectors in accuracy.

Extensive experiments validate FairOD's fairness and effectiveness across datasets.

Abstract

Fairness and Outlier Detection (OD) are closely related, as it is exactly the goal of OD to spot rare, minority samples in a given population. However, when being a minority (as defined by protected variables, such as race/ethnicity/sex/age) does not reflect positive-class membership (such as criminal/fraud), OD produces unjust outcomes. Surprisingly, fairness-aware OD has been almost untouched in prior work, as fair machine learning literature mainly focuses on supervised settings. Our work aims to bridge this gap. Specifically, we develop desiderata capturing well-motivated fairness criteria for OD, and systematically formalize the fair OD problem. Further, guided by our desiderata, we propose FairOD, a fairness-aware outlier detector that has the following desirable properties: FairOD (1) exhibits treatment parity at test time, (2) aims to flag equal proportions of samples from all groups (i.e. obtain group fairness, via statistical parity), and (3) strives to flag truly high-risk samples within each group. Extensive experiments on a diverse set of synthetic and real world datasets show that FairOD produces outcomes that are fair with respect to protected variables, while performing comparable to (and in some cases, even better than) fairness-agnostic detectors in terms of detection performance.