Jour Fixe - Automatic Location of Unfair Disparities in Machine Learning
Moritz von Zahn
Automatic Location of Unfair Disparities in Machine Learning
Machine learning was repeatedly proven to provide predictions with disparate outcomes, in which subgroups of the population (e. g., defined by age, gender, or other sensitive attributes) are systematically disadvantaged. In order to automatically locate the subgroups affected, we propose a data-driven framework that builds upon conditional inference forests and statistical hypotheses testing. Our framework implements different notions of fairness, handles both categorical and continuous attributes, and also locates “hidden” disparities due to intersectionality. We thus support practitioners in conducting algorithmic audits and protecting individuals from discrimination.