Operationalizing Individual Fairness with Pairwise Fair Representations
This addresses fairness in machine learning for applications like recidivism and crime prediction, offering a practical alternative to human-specified metrics, though it builds incrementally on existing fairness concepts.
The paper tackles the challenge of operationalizing individual fairness without requiring a human-specified similarity metric by proposing Pairwise Fair Representations (PFR) that use side-information from fairness graphs, and demonstrates its viability on real-world datasets like COMPAS and Crime & Communities.
We revisit the notion of individual fairness proposed by Dwork et al. A central challenge in operationalizing their approach is the difficulty in eliciting a human specification of a similarity metric. In this paper, we propose an operationalization of individual fairness that does not rely on a human specification of a distance metric. Instead, we propose novel approaches to elicit and leverage side-information on equally deserving individuals to counter subordination between social groups. We model this knowledge as a fairness graph, and learn a unified Pairwise Fair Representation (PFR) of the data that captures both data-driven similarity between individuals and the pairwise side-information in fairness graph. We elicit fairness judgments from a variety of sources, including human judgments for two real-world datasets on recidivism prediction (COMPAS) and violent neighborhood prediction (Crime & Communities). Our experiments show that the PFR model for operationalizing individual fairness is practically viable.