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. 2110.05000
90
43
v1v2 (latest)

Exact Matching of Random Graphs with Constant Correlation

11 October 2021
Cheng Mao
M. Rudelson
K. Tikhomirov
ArXiv (abs)PDFHTML
Abstract

This paper deals with the problem of graph matching or network alignment for Erd\H{o}s--R\'enyi graphs, which can be viewed as a noisy average-case version of the graph isomorphism problem. Let GGG and G′G'G′ be G(n,p)G(n, p)G(n,p) Erd\H{o}s--R\'enyi graphs marginally, identified with their adjacency matrices. Assume that GGG and G′G'G′ are correlated such that E[GijGij′]=p(1−α)\mathbb{E}[G_{ij} G'_{ij}] = p(1-\alpha)E[Gij​Gij′​]=p(1−α). For a permutation π\piπ representing a latent matching between the vertices of GGG and G′G'G′, denote by GπG^\piGπ the graph obtained from permuting the vertices of GGG by π\piπ. Observing GπG^\piGπ and G′G'G′, we aim to recover the matching π\piπ. In this work, we show that for every ε∈(0,1]\varepsilon \in (0,1]ε∈(0,1], there is n0>0n_0>0n0​>0 depending on ε\varepsilonε and absolute constants α0,R>0\alpha_0, R > 0α0​,R>0 with the following property. Let n≥n0n \ge n_0n≥n0​, (1+ε)log⁡n≤np≤n1Rlog⁡log⁡n(1+\varepsilon) \log n \le np \le n^{\frac{1}{R \log \log n}}(1+ε)logn≤np≤nRloglogn1​, and 0<α<min⁡(α0,ε/4)0 < \alpha < \min(\alpha_0,\varepsilon/4)0<α<min(α0​,ε/4). There is a polynomial-time algorithm FFF such that P{F(Gπ,G′)=π}=1−o(1)\mathbb{P}\{F(G^\pi,G')=\pi\}=1-o(1)P{F(Gπ,G′)=π}=1−o(1). This is the first polynomial-time algorithm that recovers the exact matching between vertices of correlated Erd\H{o}s--R\'enyi graphs with constant correlation with high probability. The algorithm is based on comparison of partition trees associated with the graph vertices.

View on arXiv
Comments on this paper