Abstract
Indirect encoding schemes for neural network phenotypes can represent large networks compactly. In previous work, we presented a new approach where networks are encoded indirectly as a set of Fourier-type coefficients that decorrelate weight matrices such that they can often be represented by a small number of genes, effectively reducing the search space dimensionality, and speed up search. Up to now, the complexity of networks using this encoding was fixed a priori, both in terms of (1) the number of free parameters (topology) and (2) the number of coefficients. In this paper, we introduce a method, called Compressed Network Complexity Search (CNCS), for automatically determining network complexity that favors parsimonious solutions. CNCS maintains a probability distribution over complexity classes that it uses to select which class to optimize. Class probabilities are adapted based on their expected fitness. Starting with a prior biased toward the simplest networks, the distribution grows gradually until a solution is found. Experiments on two benchmark control problems, including a challenging non-linear version of the helicopter hovering task, demonstrate that the method consistently finds simple solutions.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Abbeel, P., Ganapathi, V., Ng, A.Y.: Learning vehicular dynamics, with application to modeling helicopters. In: NIPS (2005)
Dürr, P., Mattiussi, C., Floreano, D.: Neuroevolution with Analog Genetic Encoding. In: Runarsson, T.P., Beyer, H.-G., Burke, E.K., Merelo-Guervós, J.J., Whitley, L.D., Yao, X. (eds.) PPSN 2006. LNCS, vol. 4193, pp. 671–680. Springer, Heidelberg (2006)
Gruau, F.: Cellular encoding of genetic neural networks. Technical Report RR-92-21, Ecole Normale Superieure de Lyon, Institut IMAG, Lyon, France (1992)
Kitano, H.: Designing neural networks using genetic algorithms with graph generation system. Complex Systems 4, 461–476 (1990)
Koutník, J., Gomez, F., Schmidhuber, J.: Evolving neural networks in compressed weight space. In: Proceedings of the Conference on Genetic and Evolutionary Computation, GECCO 2010 (2010)
Koutník, J., Gomez, F., Schmidhuber, J.: Searching for minimal neural networks in fourier space. In: Proc. of the 4th Conf. on Artificial General Intelligence (2010)
Levin, L.A.: Universal sequential search problems. Problems of Information Transmission 9(3), 265–266 (1973)
Parzen, E.: On estimation of a probability density function and mode. The Annals of Mathematical Statistics 33(3), 1065–1076 (1962)
Schmidhuber, J.: Discovering neural nets with low Kolmogorov complexity and high generalization capability. Neural Networks 10(5), 857–873 (1997)
Stanley, K.O., Miikkulainen, R.: Evolving neural networks through augmenting topologies. Evolutionary Computation 10, 99–127 (2002)
Stanley, K.O., Miikkulainen, R.: A taxonomy for artificial embryogeny. Artificial Life 9(2), 93–130 (2003)
Wierstra, D., Schaul, T., Peters, J., Schmidhuber, J.: Natural Evolution Strategies. In: Proceedings of the Congress on Evolutionary Computation (CEC 2008), Hongkong. IEEE Press (2008)
Wierstra, D., Schaul, T., Sun, T.G.Y., Schmidhuber, J.: Natural evolution strategies. Technical report (2011), arXiv:1106.4487v1
Woolley, B.G., Stanley, K.O.: Evolving a Single Scalable Controller for an Octopus Arm with a Variable Number of Segments. In: Schaefer, R., Cotta, C., Kołodziej, J., Rudolph, G. (eds.) PPSN XI. LNCS, vol. 6239, pp. 270–279. Springer, Heidelberg (2010)
Zhang, B.-T., Muhlenbein, H.: Evolving optimal neural networks using genetic algorithms with Occam’s razor. Complex Systems 7, 199–220 (1993)
Zhang, B.-T., Muhlenbein, H.: Balancing accuracy and parsimony in genetic programming. Evolutionary Computation 3, 17–38 (1995)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2012 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Gomez, F., Koutník, J., Schmidhuber, J. (2012). Compressed Network Complexity Search. In: Coello, C.A.C., Cutello, V., Deb, K., Forrest, S., Nicosia, G., Pavone, M. (eds) Parallel Problem Solving from Nature - PPSN XII. PPSN 2012. Lecture Notes in Computer Science, vol 7491. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-32937-1_32
Download citation
DOI: https://doi.org/10.1007/978-3-642-32937-1_32
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-32936-4
Online ISBN: 978-3-642-32937-1
eBook Packages: Computer ScienceComputer Science (R0)