TY - JOUR
T1 - Can Fairness be Automated?
T2 - Guidelines and Opportunities for Fairness-aware AutoML
AU - Weerts, Hilde
AU - Pfisterer, Florian
AU - Feurer, Matthias
AU - Eggensperger, Katharina
AU - Bergman, Edward
AU - Awad, Noor
AU - Vanschoren, Joaquin
AU - Pechenizkiy, Mykola
AU - Bischl, Bernd
AU - Hutter, Frank
PY - 2024
Y1 - 2024
N2 - The field of automated machine learning (AutoML) introduces techniques that automate parts of the development of machine learning (ML) systems, accelerating the process and reducing barriers for novices. However, decisions derived from ML models can reproduce, amplify, or even introduce unfairness in our societies, causing harm to (groups of) individuals. In response, researchers have started to propose AutoML systems that jointly optimize fairness and predictive performance to mitigate fairness-related harm. However, fairness is a complex and inherently interdisciplinary subject, and solely posing it as an optimization problem can have adverse side effects. With this work, we aim to raise awareness among developers of AutoML systems about such limitations of fairness-aware AutoML, while also calling attention to the potential of AutoML as a tool for fairness research. We present a comprehensive overview of different ways in which fairness-related harm can arise and the ensuing implications for the design of fairness-aware AutoML. We conclude that while fairness cannot be automated, fairness-aware AutoML can play an important role in the toolbox of ML practitioners. We highlight several open technical challenges for future work in this direction. Additionally, we advocate for the creation of more user-centered assistive systems designed to tackle challenges encountered in fairness work.
AB - The field of automated machine learning (AutoML) introduces techniques that automate parts of the development of machine learning (ML) systems, accelerating the process and reducing barriers for novices. However, decisions derived from ML models can reproduce, amplify, or even introduce unfairness in our societies, causing harm to (groups of) individuals. In response, researchers have started to propose AutoML systems that jointly optimize fairness and predictive performance to mitigate fairness-related harm. However, fairness is a complex and inherently interdisciplinary subject, and solely posing it as an optimization problem can have adverse side effects. With this work, we aim to raise awareness among developers of AutoML systems about such limitations of fairness-aware AutoML, while also calling attention to the potential of AutoML as a tool for fairness research. We present a comprehensive overview of different ways in which fairness-related harm can arise and the ensuing implications for the design of fairness-aware AutoML. We conclude that while fairness cannot be automated, fairness-aware AutoML can play an important role in the toolbox of ML practitioners. We highlight several open technical challenges for future work in this direction. Additionally, we advocate for the creation of more user-centered assistive systems designed to tackle challenges encountered in fairness work.
UR - http://www.scopus.com/inward/record.url?scp=85187994896&partnerID=8YFLogxK
U2 - 10.1613/jair.1.14747
DO - 10.1613/jair.1.14747
M3 - Article
AN - SCOPUS:85187994896
SN - 1076-9757
VL - 79
SP - 640
EP - 677
JO - Journal of Artificial Intelligence Research
JF - Journal of Artificial Intelligence Research
ER -