A Neural Computation paper on graph isomorphism
Marcello Pelillo
pelillo at dsi.unive.it
Wed Dec 16 09:50:19 EST 1998
The following paper, accepted for publication in Neural Computation,
is accessible at the following www site:
http://www.dsi.unive.it/~pelillo/papers/nc98.ps.gz
A shorter version of it has just been presented at NIPS*98, and can be
accesses at:
http://www.dsi.unive.it/~pelillo/papers/nips.ps.gz
(files are gzipped postscripts)
Comments and suggestions are welcome!
Best regards,
Marcello Pelillo
========================
Replicator Equations, Maximal Cliques, and Graph Isomorphism
Marcello Pelillo
University of Venice, Italy
ABSTRACT
We present a new energy-minimization framework for the graph isomorphism
problem which is based on an equivalent maximum clique formulation. The
approach is centered around a fundamental result proved by Motzkin and
Straus in the mid-1960s, and recently expanded in various ways, which
allows us to formulate the maximum clique problem in terms of a standard
quadratic program. The attractive feature of this formulation is that a
clear one-to-one correspondence exists between the solutions of the
quadratic program and those in the original, combinatorial problem. To
solve the program we use the so-called ``replicator'' equations, a class
of straightforward continuous- and discrete-time dynamical systems
developed in various branches of theoretical biology. We show how,
despite their inherent inability to escape from local solutions, they
nevertheless provide experimental results which are competitive with those
obtained using more elaborate mean-field annealing heuristics.
====================
________________________________________________________________________
Marcello Pelillo
Dipartimento di Informatica
Universita' Ca' Foscari di Venezia
Via Torino 155, 30172 Venezia Mestre, Italy
Tel: (39) 41 2908.440
Fax: (39) 41 2908.419
E-mail: pelillo at dsi.unive.it
URL: http://www.dsi.unive.it/~pelillo
More information about the Connectionists
mailing list