Achieving Fairness via Post-Processing in Web-Scale Recommender Systems

TL;DR

This paper extends fairness definitions to recommender systems, proposing scalable, model-agnostic post-processing methods to ensure fairness measures like equality of opportunity and equalized odds, effectively addressing position bias.

Contribution

It introduces scalable, model-agnostic post-processing algorithms for fairness in rankings, specifically targeting position bias in web-scale recommender systems.

Findings

Algorithms effectively improve fairness metrics in simulations.

Real-world experiments confirm practical applicability.

Methods are scalable and model-agnostic.

Abstract

Building fair recommender systems is a challenging and crucial area of study due to its immense impact on society. We extended the definitions of two commonly accepted notions of fairness to recommender systems, namely equality of opportunity and equalized odds. These fairness measures ensure that equally "qualified" (or "unqualified") candidates are treated equally regardless of their protected attribute status (such as gender or race). We propose scalable methods for achieving equality of opportunity and equalized odds in rankings in the presence of position bias, which commonly plagues data generated from recommender systems. Our algorithms are model agnostic in the sense that they depend only on the final scores provided by a model, making them easily applicable to virtually all web-scale recommender systems. We conduct extensive simulations as well as real-world experiments to show the efficacy of our approach.