GRAPH MATCHING VIA CONVEX RELAXATION TO THE SIMPLEX

被引:0
|
作者
Araya, Ernesto [1 ]
Tyagi, Hemant [2 ]
机构
[1] Ludwig Maximilians Univ Munchen, Dept Math, Munich, Germany
[2] Univ Lille, Inria, CNRS, UMR 8524,Lab Paul Painleve, F-59000 Lille, France
来源
FOUNDATIONS OF DATA SCIENCE | 2024年
关键词
Graph matching; machine learning; mirror descent; multiplicative weight updates; PROTEIN-INTERACTION NETWORKS; GLOBAL ALIGNMENT; ALGORITHMS; DESCENT;
D O I
10.3934/fods.2024034
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
This paper addresses the Graph Matching problem, which consists of finding the best possible alignment between two input graphs, and has many applications in computer vision, network deanonymization and protein alignment. A common approach to tackle this problem is through convex relaxations Here, we introduce a new convex relaxation onto the unit simplex and develop an efficient mirror descent scheme with closed-form iterations for solving this problem. Under the correlated Gaussian Wigner model, we show that the simplex relaxation admits a unique solution with high probability. In the noiseless case, this is shown to imply exact recovery of the ground truth permutation. Additionally, we establish a novel sufficiency condition for the input matrix in standard greedy rounding methods, which is less restrictive than the commonly used 'diagonal dominance' condition. We use this condition to show exact one-step recovery of the ground truth (holding almost surely) via the mirror descent scheme, in the noiseless setting. We also use this condition to obtain significantly improved conditions for the GRAMPA algorithm [26] in the noiseless setting. Our method is evaluated on both synthetic and real data, demonstrating superior statistical performance compared to existing convex relaxation methods with similar computational costs.
引用
收藏
页码:464 / 501
页数:38
相关论文
共 50 条
  • [1] A graph matching algorithm based on concavely regularized convex relaxation
    Liu, Zhi-Yong
    Qiao, Hong
    Jia, Li-Hao
    Xu, Lei
    NEUROCOMPUTING, 2014, 134 : 140 - 148
  • [2] Graph Matching by Simplified Convex-Concave Relaxation Procedure
    Zhi-Yong Liu
    Hong Qiao
    Xu Yang
    Steven C. H. Hoi
    International Journal of Computer Vision, 2014, 109 : 169 - 186
  • [3] Graph Matching by Simplified Convex-Concave Relaxation Procedure
    Liu, Zhi-Yong
    Qiao, Hong
    Yang, Xu
    Hoi, Steven C. H.
    INTERNATIONAL JOURNAL OF COMPUTER VISION, 2014, 109 (03) : 169 - 186
  • [4] On convex relaxation of graph isomorphism
    Aflalo, Yonathan
    Bronstein, Alexander
    Kimmel, Ron
    PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2015, 112 (10) : 2942 - 2947
  • [5] Graph matching by neural relaxation
    Turner, M
    Austin, J
    NEURAL COMPUTING & APPLICATIONS, 1998, 7 (03) : 238 - 248
  • [6] Lagrangian relaxation graph matching
    Jiang, Bo
    Tang, Jin
    Cao, Xiaochun
    Luo, Bin
    PATTERN RECOGNITION, 2017, 61 : 255 - 265
  • [7] Graph matching by neural relaxation
    M. Turner
    J. Austin
    Neural Computing & Applications, 1998, 7 : 238 - 248
  • [8] Graph matching with hierarchical discrete relaxation
    Wilson, RC
    Hancock, ER
    PATTERN RECOGNITION LETTERS, 1999, 20 (10) : 1041 - 1052
  • [9] Graph matching by relaxation of fuzzy assignments
    Medasani, S
    Krishnapuram, R
    Choi, Y
    IEEE TRANSACTIONS ON FUZZY SYSTEMS, 2001, 9 (01) : 173 - 182
  • [10] A new relaxation model for weighted graph matching
    Zheng K.-J.
    Gao Y.-T.
    Peng J.-G.
    Zidonghua Xuebao/Acta Automatica Sinica, 2010, 36 (08): : 1200 - 1203