ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2202.01391
63
9

Fair Representation Clustering with Several Protected Classes

3 February 2022
Zhen Dai
Yury Makarychev
A. Vakilian
    FaML
ArXivPDFHTML
Abstract

We study the problem of fair kkk-median where each cluster is required to have a fair representation of individuals from different groups. In the fair representation kkk-median problem, we are given a set of points XXX in a metric space. Each point x∈Xx\in Xx∈X belongs to one of ℓ\ellℓ groups. Further, we are given fair representation parameters αj\alpha_jαj​ and βj\beta_jβj​ for each group j∈[ℓ]j\in [\ell]j∈[ℓ]. We say that a kkk-clustering C1,⋯ ,CkC_1, \cdots, C_kC1​,⋯,Ck​ fairly represents all groups if the number of points from group jjj in cluster CiC_iCi​ is between αj∣Ci∣\alpha_j |C_i|αj​∣Ci​∣ and βj∣Ci∣\beta_j |C_i|βj​∣Ci​∣ for every j∈[ℓ]j\in[\ell]j∈[ℓ] and i∈[k]i\in [k]i∈[k]. The goal is to find a set C\mathcal{C}C of kkk centers and an assignment ϕ:X→C\phi: X\rightarrow \mathcal{C}ϕ:X→C such that the clustering defined by (C,ϕ)(\mathcal{C}, \phi)(C,ϕ) fairly represents all groups and minimizes the ℓ1\ell_1ℓ1​-objective ∑x∈Xd(x,ϕ(x))\sum_{x\in X} d(x, \phi(x))∑x∈X​d(x,ϕ(x)). We present an O(log⁡k)O(\log k)O(logk)-approximation algorithm that runs in time nO(ℓ)n^{O(\ell)}nO(ℓ). Note that the known algorithms for the problem either (i) violate the fairness constraints by an additive term or (ii) run in time that is exponential in both kkk and ℓ\ellℓ. We also consider an important special case of the problem where αj=βj=fjf\alpha_j = \beta_j = \frac{f_j}{f}αj​=βj​=ffj​​ and fj,f∈Nf_j, f \in \mathbb{N}fj​,f∈N for all j∈[ℓ]j\in [\ell]j∈[ℓ]. For this special case, we present an O(log⁡k)O(\log k)O(logk)-approximation algorithm that runs in (kf)O(ℓ)log⁡n+poly(n)(kf)^{O(\ell)}\log n + poly(n)(kf)O(ℓ)logn+poly(n) time.

View on arXiv
Comments on this paper