Manifold embedding of graphs using the heat kernel

Authors:
Xiao Bai;Richard C. Wilson;Edwin R. Hancock
Affiliations:
Department of Computer Science, University of York, York, UK;Department of Computer Science, University of York, York, UK;Department of Computer Science, University of York, York, UK
Venue:
IMA'05 Proceedings of the 11th IMA international conference on Mathematics of Surfaces
Year:
2005

Citing 5
Cited 2

An Eigendecomposition Approach to Weighted Graph Matching Problems

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Spectral Algorithm for Seriation and the Consecutive Ones Problem

SIAM Journal on Computing
Normalized Cuts and Image Segmentation

IEEE Transactions on Pattern Analysis and Machine Intelligence
Structural Graph Matching Using the EM Algorithm and Singular Value Decomposition

IEEE Transactions on Pattern Analysis and Machine Intelligence - Graph Algorithms and Computer Vision
Advances in Large Margin Classifiers

Advances in Large Margin Classifiers

Recognising Facial Expressions Using Spherical Harmonics

SSPR & SPR '08 Proceedings of the 2008 Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition
Invariants of distance k-graphs for graph embedding

Pattern Recognition Letters

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we investigate the use of heat kernels as a means of embedding the individual nodes of a graph on a manifold in a vector space. The heat kernel of the graph is found by exponentiating the Laplacian eigen-system over time. We show how the spectral representation of the heat kernel can be used to compute both Euclidean and geodesic distances between nodes. We use the resulting pattern of distances to embed the nodes of the graph on a manifold using multidimensional scaling. The distribution of embedded points can be used to characterise the graph, and can be used for the purposes of graph clustering. Here for the sake of simplicity, we use spatial moments. We experiment with the resulting algorithms on the COIL database, and they are demonstrated to offer a useful margin of advantage over existing alternatives.