Fairness in representation: quantifying stereotyping as a representational harm
It addresses the understudied problem of representational harms in algorithmic fairness for ML practitioners and researchers, offering incremental contributions.
The paper formalizes two notions of stereotyping as representational harms in machine learning and shows how they lead to allocative harms, proposing mitigation strategies that are demonstrated on synthetic datasets.
While harms of allocation have been increasingly studied as part of the subfield of algorithmic fairness, harms of representation have received considerably less attention. In this paper, we formalize two notions of stereotyping and show how they manifest in later allocative harms within the machine learning pipeline. We also propose mitigation strategies and demonstrate their effectiveness on synthetic datasets.