Addressing Polarization and Unfairness in Performative Prediction

TL;DR

This paper investigates the societal impacts of performative prediction, revealing issues of polarization and unfairness, and introduces new fairness mechanisms that ensure stability and fairness in models influenced by feedback loops.

Contribution

The paper identifies fairness challenges in performative prediction and proposes novel mechanisms that guarantee both stability and fairness, supported by theoretical and empirical validation.

Findings

PS solutions can cause polarization and disparities

Conventional fairness methods often fail under distribution shifts

Proposed mechanisms ensure stability and fairness in PP models

Abstract

In many real-world applications of machine learning such as recommendations, hiring, and lending, deployed models influence the data they are trained on, leading to feedback loops between predictions and data distribution. The performative prediction (PP) framework captures this phenomenon by modeling the data distribution as a function of the deployed model. While prior work has focused on finding performative stable (PS) solutions for robustness, their societal impacts, particularly regarding fairness, remain underexplored. We show that PS solutions can lead to severe polarization and prediction performance disparities, and that conventional fairness interventions in previous works often fail under model-dependent distribution shifts due to failing the PS criteria. To address these challenges in PP, we introduce novel fairness mechanisms that provably ensure both stability and fairness, validated by theoretical analysis and empirical results.