Preprocessing Matters: Automated Pipeline Selection for Fair Classification

González-Zelaya, Vladimiro; Salas, Julián; Prangle, Dennis; Missier, Paolo

doi:10.1007/978-3-031-33498-6_14

Preprocessing Matters: Automated Pipeline Selection for Fair Classification

Journal

Modeling Decisions for Artificial Intelligence

Lecture Notes in Computer Science

ISSN

0302-9743

1611-3349

Date Issued

2023

Author(s)

González-Zelaya, Vladimiro

Facultad de Ciencias Económicas y Empresariales - CampCM

Salas, Julián

Prangle, Dennis

Missier, Paolo

Type

text::book::book part

DOI

10.1007/978-3-031-33498-6_14

URL

https://scripta.up.edu.mx/handle/20.500.12552/3688

Abstract

Improving fairness by manipulating the preprocessing stages of classification pipelines is an active area of research, closely related to AutoML. We propose a genetic optimisation algorithm, FairPipes, which optimises for user-defined combinations of fairness and accuracy and for multiple definitions of fairness, providing flexibility in the fairness-accuracy trade-off. FairPipes heuristically searches through a large space of pipeline configurations, achieving near-optimality efficiently, presenting the user with an estimate of the solutions’ Pareto front. We also observe that the optimal pipelines differ for different datasets, suggesting that no “universal best” pipeline exists and confirming that FairPipes fills a niche in the fairness-aware AutoML space.