English

Communication-Optimal Distributed Dynamic Graph Clustering

Data Structures and Algorithms 2018-11-16 v1 Computational Complexity

Abstract

We consider the problem of clustering graph nodes over large-scale dynamic graphs, such as citation networks, images and web networks, when graph updates such as node/edge insertions/deletions are observed distributively. We propose communication-efficient algorithms for two well-established communication models namely the message passing and the blackboard models. Given a graph with nn nodes that is observed at ss remote sites over time [1,t][1,t], the two proposed algorithms have communication costs O~(ns)\tilde{O}(ns) and O~(n+s)\tilde{O}(n+s) (O~\tilde{O} hides a polylogarithmic factor), almost matching their lower bounds, Ω(ns)\Omega(ns) and Ω(n+s)\Omega(n+s), respectively, in the message passing and the blackboard models. More importantly, we prove that at each time point in [1,t][1,t] our algorithms generate clustering quality nearly as good as that of centralizing all updates up to that time and then applying a standard centralized clustering algorithm. We conducted extensive experiments on both synthetic and real-life datasets which confirmed the communication efficiency of our approach over baseline algorithms while achieving comparable clustering results.

Keywords

Cite

@article{arxiv.1811.06072,
  title  = {Communication-Optimal Distributed Dynamic Graph Clustering},
  author = {Chun Jiang Zhu and Tan Zhu and Kam-Yiu Lam and Song Han and Jinbo Bi},
  journal= {arXiv preprint arXiv:1811.06072},
  year   = {2018}
}

Comments

Accepted and to appear in AAAI'19

R2 v1 2026-06-23T05:16:03.885Z