Welfare-Centric Clustering

2025-08-14

AI Summary

Traditional fair clustering often yields counterintuitive outcomes by prioritizing group representation or cost balance while neglecting collective group welfare. This paper introduces a welfare-centered clustering paradigm, where group utility—jointly modeling individual distance costs and proportional representation—is the primary optimization objective. We are the first to incorporate Rawlsian (maximizing the utility of the worst-off group) and Utilitarian (maximizing total group utility) social welfare functions as formal fairness criteria. To optimize these objectives, we design combinatorial algorithms with theoretical guarantees—including exact and approximation algorithms—with provable performance bounds. Extensive experiments on multiple real-world datasets demonstrate that our approach significantly improves both inter-group utility equity and aggregate welfare, consistently outperforming state-of-the-art fair clustering methods in quantitative and qualitative evaluations.

Technology Category

Machine Learning: ClusteringGame Theory and Economic Paradigms: Fair DivisionConstraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

User Modeling, Personalization and Recommendation: Fairness-aware retrieval and rankingEconomics, Online Markets and Human Computation: Fairness and ethical considerations in crowd work and in human-in-the-loop AI systemsWeb Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text

Abstract

Fair clustering has traditionally focused on ensuring equitable group representation or equalizing group-specific clustering costs. However, Dickerson et al. (2025) recently showed that these fairness notions may yield undesirable or unintuitive clustering outcomes and advocated for a welfare-centric clustering approach that models the utilities of the groups. In this work, we model group utilities based on both distances and proportional representation and formalize two optimization objectives based on welfare-centric clustering: the Rawlsian (Egalitarian) objective and the Utilitarian objective. We introduce novel algorithms for both objectives and prove theoretical guarantees for them. Empirical evaluations on multiple real-world datasets demonstrate that our methods significantly outperform existing fair clustering baselines.

Problem

• Modeling group utilities based on distances and representation
• Formalizing Rawlsian and Utilitarian welfare-centric clustering objectives
• Developing novel algorithms with theoretical guarantees for welfare-centric clustering

Innovation

• Welfare-centric clustering with group utilities
• Novel algorithms for Rawlsian and Utilitarian objectives
• Theoretical guarantees and empirical performance improvements