Solving inexact graph isomorphism problems using neural networks

We present a neural network approach to solve exact and inexact graph isomorphism problems for weighted graphs. In contrast to other neural heuristics or related methods this approach is based on a neural refinement procedure to reduce the search space followed by an energy-minimizing matching process. Experiments on random weighted graphs in the range of 100–5000 vertices and on chemical molecular structures are presented and discussed. r 2004 Elsevier B.V. All rights reserved.