Recommender Systems as Control Systems

TL;DR

This paper introduces a control-theoretic perspective on recommender systems to analyze how fairness interventions influence long-term behavior and system performance.

Contribution

It offers a novel control-theoretic framework to understand fairness impacts in recommender systems over time.

Findings

Fairness interventions can improve overall system performance when considering long-term dynamics.

A control-theoretic approach reveals that fairness is not necessarily a trade-off against utility.

Understanding system dynamics is crucial for effective fairness strategies.

Abstract

We propose a control-theoretic interpretation of recommender systems and use this perspective to analyze how fairness interventions shape long-term system behavior. Fairness concerns arise for both users and creators, ranging from opinion polarization and representation bias on the user side to popularity bias on the creator side. A central insight of our analysis is that fairness should not be viewed as a simple trade-off against utility. When optimized over time, it can in fact be beneficial for overall system performance. Realizing these gains, however, requires a clear understanding of the underlying dynamics.