Automorphism Partitioning with Neural Networks

Authors:
Brijnesh J. Jain;Fritz Wysotzki
Affiliations:
Department of Computer Science, Technical University Berlin, Germany. e-mail: bjj@cs.tu-berlin.de;Department of Computer Science, Technical University Berlin, Germany. e-mail: wysotzki@cs.tu-berlin.de
Venue:
Neural Processing Letters
Year:
2003

Citing 5
Cited 1

Matching Hierarchical Structures Using Association Graphs

IEEE Transactions on Pattern Analysis and Machine Intelligence
Replicator equations, maximal cliques, and graph isomorphism

Neural Computation
Fast Winner-Takes-All Networks for the Maximum Clique Problem

KI '02 Proceedings of the 25th Annual German Conference on AI: Advances in Artificial Intelligence
Neural Networks for Combinatorial Optimization: a Review of More Than a Decade of Research

INFORMS Journal on Computing
Approximating maximum clique with a Hopfield network

IEEE Transactions on Neural Networks

Towards the unification of structural and statistical pattern recognition

Pattern Recognition Letters

Quantified Score

Hi-index	0.02

Visualization

Abstract

We present a neural approach for approximating the automorphism partitioning problem of a given graph. This approach combines the energy minimization process of neural networks for combinatorial optimization problems with simple group-theoretic properties. Neural networks are applied to rapidly find relevant automorphisms while group-theoretic information guides the search for these automorphisms.