English

Approximation Algorithms for Socially Fair Clustering

Data Structures and Algorithms 2021-07-16 v2 Machine Learning Machine Learning

Abstract

We present an (eO(p)logloglog)(e^{O(p)} \frac{\log \ell}{\log\log\ell})-approximation algorithm for socially fair clustering with the p\ell_p-objective. In this problem, we are given a set of points in a metric space. Each point belongs to one (or several) of \ell groups. The goal is to find a kk-medians, kk-means, or, more generally, p\ell_p-clustering that is simultaneously good for all of the groups. More precisely, we need to find a set of kk centers CC so as to minimize the maximum over all groups jj of u in group jd(u,C)p\sum_{u \text{ in group }j} d(u,C)^p. The socially fair clustering problem was independently proposed by Ghadiri, Samadi, and Vempala [2021] and Abbasi, Bhaskara, and Venkatasubramanian [2021]. Our algorithm improves and generalizes their O()O(\ell)-approximation algorithms for the problem. The natural LP relaxation for the problem has an integrality gap of Ω()\Omega(\ell). In order to obtain our result, we introduce a strengthened LP relaxation and show that it has an integrality gap of Θ(logloglog)\Theta(\frac{\log \ell}{\log\log\ell}) for a fixed pp. Additionally, we present a bicriteria approximation algorithm, which generalizes the bicriteria approximation of Abbasi et al. [2021].

Keywords

Cite

@article{arxiv.2103.02512,
  title  = {Approximation Algorithms for Socially Fair Clustering},
  author = {Yury Makarychev and Ali Vakilian},
  journal= {arXiv preprint arXiv:2103.02512},
  year   = {2021}
}

Comments

COLT 2021

R2 v1 2026-06-23T23:43:05.312Z