# Operationalizing Individual Fairness with Pairwise Fair Representations

## TL;DR

This paper introduces a method to implement individual fairness without needing a human-defined similarity metric by using side-information and fairness graphs to learn fair representations, demonstrated on real-world datasets.

## Contribution

It proposes a novel approach to operationalize individual fairness without a human-specified metric, leveraging side-information and fairness graphs to learn fair data representations.

## Key findings

- The Pairwise Fair Representation (PFR) effectively captures data similarity and fairness side-information.
- The approach is practically viable on real-world datasets like COMPAS and Crime & Communities.
- Experimental results demonstrate improved fairness in predictions.

## Abstract

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.