A Catalog of Fairness-Aware Practices in Machine Learning Engineering

TL;DR

This paper addresses the challenge of integrating fairness into ML-enabled systems by compiling a catalog of 28 fairness-aware practices mapped across six stages of the ML development lifecycle through a systematic mapping study. The authors detail a rigorous methodology (Scopus-based search, inclusion/exclusion criteria, and iterative content analysis) to extract practices from 135 primary studies, distilling them into a structured catalog that spans requirements elicitation to maintenance. The results reveal that data preparation is the most emphasized phase for fairness interventions and demonstrate how the catalog can guide practitioners with concrete actions, including a COMPAS-based scenario for practical grounding. The work also provides a roadmap for future research, including empirical validation, fairness analytics tooling, and automation to support fair AI development, with broader implications for SE4AI, software quality, and empirical software engineering.

Abstract

Machine learning's widespread adoption in decision-making processes raises concerns about fairness, particularly regarding the treatment of sensitive features and potential discrimination against minorities. The software engineering community has responded by developing fairness-oriented metrics, empirical studies, and approaches. However, there remains a gap in understanding and categorizing practices for engineering fairness throughout the machine learning lifecycle. This paper presents a novel catalog of practices for addressing fairness in machine learning derived from a systematic mapping study. The study identifies and categorizes 28 practices from existing literature, mapping them onto different stages of the machine learning lifecycle. From this catalog, the authors extract actionable items and implications for both researchers and practitioners in software engineering. This work aims to provide a comprehensive resource for integrating fairness considerations into the development and deployment of machine learning systems, enhancing their reliability, accountability, and credibility.

Figures (3)

• Figure 1: Overview of the methodology to achieve the objective of the study.
• Figure 2: Overview of the Search Process Execution.
• Figure 3: Research roadmap on Fairness-aware development and recommendations.