Long-term Dynamics of Fairness Intervention in Connection Recommender Systems

TL;DR

This paper investigates the long-term effects of fairness interventions in connection recommender systems, revealing that common approaches may inadvertently amplify biases over time due to feedback loops.

Contribution

It introduces a dynamic analysis of fairness interventions in recommender systems, highlighting potential long-term bias amplification overlooked by static evaluations.

Findings

Common fairness interventions fail to prevent bias amplification over time.

Theoretical analysis using a Pólya urn model explains bias dynamics.

Long-term bias can increase despite seemingly fair aggregate outcomes.

Abstract

Recommender system fairness has been studied from the perspectives of a variety of stakeholders including content producers, the content itself and recipients of recommendations. Regardless of which type of stakeholders are considered, most works in this area assess the efficacy of fairness intervention by evaluating a single fixed fairness criterion through the lens of a one-shot, static setting. Yet recommender systems constitute dynamical systems with feedback loops from the recommendations to the underlying population distributions which could lead to unforeseen and adverse consequences if not taken into account. In this paper, we study a connection recommender system patterned after the systems employed by web-scale social networks and analyze the long-term effects of intervening on fairness in the recommendations. We find that, although seemingly fair in aggregate, common exposure and utility parity interventions fail to mitigate amplification of biases in the long term. We theoretically characterize how certain fairness interventions impact the bias amplification dynamics in a stylized P\'{o}lya urn model.