A message passing graph match algorithm based on a generative graphical model

Authors:
Gang Shen;Wei Li
Affiliations:
School of Software Engineering, Huazhong University of Science and Technology, Wuhan, China;School of Software Engineering, Huazhong University of Science and Technology, Wuhan, China
Venue:
AMT'12 Proceedings of the 8th international conference on Active Media Technology
Year:
2012

Citing 12
Cited 0

A shortest augmenting path algorithm for dense and sparse linear assignment problems

Computing
A Graduated Assignment Algorithm for Graph Matching

IEEE Transactions on Pattern Analysis and Machine Intelligence
An Algorithm for Subgraph Isomorphism

Journal of the ACM (JACM)
Similarity Flooding: A Versatile Graph Matching Algorithm and Its Application to Schema Matching

ICDE '02 Proceedings of the 18th International Conference on Data Engineering
Graphical models for graph matching: Approximate models and optimal algorithms

Pattern Recognition Letters - Special issue: In memoriam Azriel Rosenfeld
Feature-based similarity search in graph structures

ACM Transactions on Database Systems (TODS)
Learning Graph Matching

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Path Following Algorithm for the Graph Matching Problem

IEEE Transactions on Pattern Analysis and Machine Intelligence
Graph Node Clustering via Transitive Node Similarity

PCI '10 Proceedings of the 2010 14th Panhellenic Conference on Informatics
Loopy belief propagation for approximate inference: an empirical study

UAI'99 Proceedings of the Fifteenth conference on Uncertainty in artificial intelligence
Max-Product for Maximum Weight Matching: Convergence, Correctness, and LP Duality

IEEE Transactions on Information Theory
A Probabilistic Approach to Spectral Graph Matching

IEEE Transactions on Pattern Analysis and Machine Intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

In This paper, we present a generative model to measure the graph similarity, assuming that an observed graph is generated from a template by a Markov random field. The potentials of this random process are characterized by two sets of parameters: the attribute expectations specified by the the template graph, and the variances that can be learned by a maximum likelihood estimator from a collection of samples. Once a sample graph is observed, a max-product loopy belief propagation algorithm is applied to approximate the most probable explanation of the template's vertices, mapped to the sample's vertices. As demonstrated by the experiments, compared with other algorithms, the proposed approach performed better for near isomorphic graphs in the typical graph alignment and information retrieval applications.