research-article

A geometric encoding for neural network evolution

Authors:

Emmanuel Rachelson,

Dennis G. WilsonAuthors Info & Claims

GECCO '21: Proceedings of the Genetic and Evolutionary Computation Conference

Pages 919 - 927

https://doi.org/10.1145/3449639.3459361

Published: 26 June 2021 Publication History

Abstract

A major limitation to the optimization of artificial neural networks (ANN) with evolutionary methods lies in the high dimensionality of the search space, the number of weights growing quadratically with the size of the network. This leads to expensive training costs, especially in evolution strategies which rely on matrices whose sizes grow with the number of genes. We introduce a geometric encoding for neural network evolution (GENE) as a representation of ANN parameters in a smaller space that scales linearly with the number of neurons, allowing for efficient parameter search. Each neuron of the network is encoded as a point in a latent space and the weight of a connection between two neurons is computed as the distance between them. The coordinates of all neurons are then optimized with evolution strategies in a reduced search space while not limiting network fitness and possibly improving search.

References

[1]

Marc G Bellemare, Yavar Naddaf, Joel Veness, and Michael Bowling. 2013. The arcade learning environment: An evaluation platform for general agents. Journal of Artificial Intelligence Research 47 (2013), 253--279.

[2]

Jeff Bezanson, Alan Edelman, Stefan Karpinski, and Viral B. Shah. 2017. Julia: A Fresh Approach to Numerical Computing. SIAM Rev. 59, 1 (2017), 65--98. : https://doi.org/10.1137/141000671.

[3]

Andrew Brock, Theodore Lim, J. M. Ritchie, and Nick Weston. 2017. SMASH: One-Shot Model Architecture Search through HyperNetworks. arXiv:1708.05344 [cs] (Aug. 2017). http://arxiv.org/abs/1708.05344 arXiv: 1708.05344.

[4]

Patryk Chrabaszcz, Ilya Loshchilov, and Frank Hutter. 2018. Back to basics: benchmarking canonical evolution strategies for playing Atari. In Proceedings of the 27th International Joint Conference on Artificial Intelligence. 1419--1426.

[5]

J. Clune, K. O. Stanley, R. T. Pennock, and C. Ofria. 2011. On the Performance of Indirect Encoding Across the Continuum of Regularity. IEEE Transactions on Evolutionary Computation 15, 3 (June 2011), 346--367. Conference Name: IEEE Transactions on Evolutionary Computation.

Digital Library

[6]

Antoine Cully, Jeff Clune, Danesh Tarapore, and Jean-Baptiste Mouret. 2015. Robots that can adapt like animals. Nature 521, 7553 (2015), 503--507.

[7]

Will Dabney, Georg Ostrovski, David Silver, and Rémi Munos. 2018. Implicit quantile networks for distributional reinforcement learning. In International conference on machine learning. PMLR, 1096--1105.

[8]

David B. D'Ambrosio and K. Stanley. 2008. Generative encoding for multiagent learning. In GECCO '08.

Digital Library

[9]

Jean Disset, Dennis G Wilson, Sylvain Cussat-Blanc, Stéphane Sanchez, Hervé Luga, and Yves Duthen. 2017. A Comparison of Genetic Regulatory Network Dynamics and Encoding. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO '17). Association for Computing Machinery, New York, NY, USA, 91--98.

Digital Library

[10]

David B D'Ambrosio, Jason Gauci, and Kenneth O Stanley. 2014. HyperNEAT: The first five years. Growing adaptive machines (2014), 159--185.

[11]

Chrisantha Fernando, Dylan Banarse, Malcolm Reynolds, Frederic Besse, David Pfau, Max Jaderberg, Marc Lanctot, and Daan Wierstra. 2016. Convolution by Evolution: Differentiable Pattern Producing Networks. arXiv:1606.02580 [cs] (June 2016). http://arxiv.org/abs/1606.02580 arXiv: 1606.02580.

[12]

Faustino Gomez, Jürgen Schmidhuber, Risto Miikkulainen, and Melanie Mitchell. 2008. Accelerated Neural Evolution through Cooperatively Coevolved Synapses. Journal of Machine Learning Research 9, 5 (2008).

[13]

David Ha. 2017. Evolving Stable Strategies. blog.otoro.net (2017). http://blog.otoro.net/2017/11/12/evolving-stable-strategies/

[14]

David Ha, Andrew M. Dai, and Quoc V. Le. 2016. HyperNetworks. (Oct. 2016). https://openreview.net/forum?id=rkpACe1lx

[15]

Nikolaus Hansen, Youhei Akimoto, and Petr Baudis. 2019. CMA-ES/pycma on Github. Published: Zenodo.

[16]

Nikolaus Hansen and Andreas Ostermeier. 2001. Completely Derandomized Self-Adaptation in Evolution Strategies. Evolutionary Computation 9, 2 (June 2001), 159--195.

Digital Library

[17]

Matthew Hausknecht, Joel Lehman, Risto Miikkulainen, and Peter Stone. 2014. A neuroevolution approach to general atari game playing. IEEE Transactions on Computational Intelligence and AI in Games 6, 4 (2014), 355--366.

[18]

Maryam Mahsal Khan, Arbab Masood Ahmad, Gul Muhammad Khan, and Julian F Miller. 2013. Fast learning neural networks using cartesian genetic programming. Neurocomputing 121 (2013), 274--289.

Digital Library

[19]

Hiroaki Kitano. 1990. Designing neural networks using genetic algorithms with graph generation system. Complex systems 4 (1990), 461--476.

[20]

Jan Koutnik, Faustino Gomez, and Jürgen Schmidhuber. 2010. Evolving neural networks in compressed weight space. In Proceedings of the 12th annual conference on Genetic and evolutionary computation. 619--626.

Digital Library

[21]

Yuqiao Liu, Yanan Sun, Bing Xue, Mengjie Zhang, and Gary Yen. 2020. A Survey on Evolutionary Neural Architecture Search. arXiv:2008.10937 [cs] (Aug. 2020). http://arxiv.org/abs/2008.10937 arXiv: 2008.10937.

[22]

H. B. Mann and D. R. Whitney. 1947. On a Test of Whether one of Two Random Variables is Stochastically Larger than the Other. The Annals of Mathematical Statistics 18, 1 (1947), 50 -- 60.

[23]

Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, and others. 2015. Human-level control through deep reinforcement learning. nature 518, 7540 (2015), 529--533. Publisher: Nature Publishing Group.

[24]

David J Montana, Lawrence Davis, and Mouiton St. 1989. Training Feedforward Neural Networks Using Genetic Algorithms. (Aug. 1989), 6.

[25]

Vinod Nair and Geoffrey E Hinton. [n. d.]. Rectified Linear Units Improve Restricted Boltzmann Machines. ([n. d.]), 8.

[26]

Martin L Puterman. 1994. Markov decision processes: discrete stochastic dynamic programming. John Wiley & Sons.

[27]

Ingo Rechenberg. 1989. Evolution Strategy: Nature's Way of Optimization. In Optimization: Methods and Applications, Possibilities and Limitations, H. W. Bergmann (Ed.). Springer Berlin Heidelberg, Berlin, Heidelberg, 106--126.

[28]

Sebastian Risi, Joel Lehman, and Kenneth O Stanley. 2010. Evolving the placement and density of neurons in the hyperneat substrate. In Proceedings of the 12th annual conference on Genetic and evolutionary computation. 563--570.

Digital Library

[29]

Sebastian Risi and Kenneth O Stanley. 2011. Enhancing es-hyperneat to evolve more complex regular neural networks. In Proceedings of the 13th annual conference on Genetic and evolutionary computation. 1539--1546.

Digital Library

[30]

Herbert Robbins and Sutton Monro. 1951. A Stochastic Approximation Method. Annals of Mathematical Statistics 22, 3 (1951), 400--407. Publisher: The Institute of Mathematical Statistics.

[31]

Edmund Ronald and Marc Schoenauer. 1994. Genetic lander: An experiment in accurate neuro-genetic control. In Parallel Problem Solving from Nature --- PPSN III. Vol. 866. Springer Berlin Heidelberg, Berlin, Heidelberg, 452--461. Series Title: Lecture Notes in Computer Science.

[32]

Kenneth O. Stanley. 2007. Compositional pattern producing networks: A novel abstraction of development. (2007).

[33]

Kenneth O Stanley, David D'Ambrosio, and Jason Gauci. 2009. A Hypercube-Based Indirect Encoding for Evolving Large-Scale Neural Networks. (2009), 39.

[34]

Kenneth O Stanley and Risto Miikkulainen. 2002. Evolving neural networks through augmenting topologies. Evolutionary computation 10, 2 (2002), 99--127. Publisher: MIT Press.

[35]

Felipe Petroski Such, Vashisht Madhavan, Edoardo Conti, Joel Lehman, Kenneth O. Stanley, and Jeff Clune. 2018. Deep Neuroevolution: Genetic Algorithms Are a Competitive Alternative for Training Deep Neural Networks for Reinforcement Learning. (2018). arXiv:cs.NE/1712.06567

[36]

Sjoerd van Steenkiste, Jan Koutník, Kurt Driessens, and Schmidhuber, Jurgen. 2016. A Wavelet-based Encoding for Neuroevolution. (2016), 8.

[37]

Darrell Whitley, Frédéric Gruau, and Larry Pyeatt. 1970. Cellular Encoding Applied to Neurocontrol. (Feb. 1970).

[38]

Daan Wierstra, Tom Schaul, Jan Peters, and Juergen Schmidhuber. 2008. Natural Evolution Strategies. In 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence). IEEE, Hong Kong, China, 3381--3387.

Cited By

Kunze TTemplier PWilson D(2024)Searching Search Spaces: Meta-evolving a Geometric Encoding for Neural Networks2024 IEEE Congress on Evolutionary Computation (CEC)10.1109/CEC60901.2024.10612026(1-8)Online publication date: 30-Jun-2024
https://doi.org/10.1109/CEC60901.2024.10612026
Baratchi MWang CLimmer Svan Rijn JHoos HBäck TOlhofer M(2024)Automated machine learning: past, present and futureArtificial Intelligence Review10.1007/s10462-024-10726-157:5Online publication date: 18-Apr-2024
https://doi.org/10.1007/s10462-024-10726-1
Papp LHaberl DEcsedi BSpielvogel CKrajnc DGrahovac MMoradi SDrexler W(2023)DEBI-NNNeural Networks10.1016/j.neunet.2023.08.026167:C(517-532)Online publication date: 1-Oct-2023
https://dl.acm.org/doi/10.1016/j.neunet.2023.08.026

Index Terms

A geometric encoding for neural network evolution
1. Computing methodologies
  1. Machine learning
    1. Machine learning approaches
      1. Bio-inspired approaches
        Generative and developmental approaches
        Genetic algorithms
      2. Neural networks

Recommendations

An enhanced hypercube-based encoding for evolving the placement, density, and connectivity of neurons

Intelligence in nature is the product of living brains, which are themselves the product of natural evolution. Although researchers in the field of neuroevolution NE attempt to recapitulate this process, artificial neural networks ANNs so far evolved ...
GMDH-type neural network algorithm with a feedback loop for structural identification of RBF neural network

In this paper, a Group Method of Data Handling (GMDH)-type neural network algorithm with a feedback loop for structural identification of Radial Basis Function (RBF) neural network is proposed. In case of the GMDH-type neural network, the network ...
Towards designing artificial neural networks by evolution
Special issue on articficial life and robotics

Designing artificial neural networks (ANNs) for different applications has been a key issue in the ANN field. At present, ANN design still relies heavily on human experts who have sufficient knowledge about ANNs and the problem to be solved. As ANN ...

Comments

Please enable JavaScript to view thecomments powered by Disqus.

Information & Contributors

Information

Published In

cover image ACM Conferences

GECCO '21: Proceedings of the Genetic and Evolutionary Computation Conference

June 2021

1219 pages

ISBN:9781450383509

DOI:10.1145/3449639

Editor:
Francisco Chicano
University of Malaga
,
General Chair:
Krzysztof Krawiec
Poznan University of Technology

Copyright © 2021 ACM.

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]

Sponsors

SIGEVO: ACM Special Interest Group on Genetic and Evolutionary Computation

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 26 June 2021

Permissions

Request permissions for this article.

Request Permissions

Check for updates

Author Tags

Qualifiers

Research-article

Conference

GECCO '21

Sponsor:

SIGEVO

GECCO '21: Genetic and Evolutionary Computation Conference

July 10 - 14, 2021

Lille, France

Acceptance Rates

Overall Acceptance Rate 1,669 of 4,410 submissions, 38%

Contributors

Other Metrics

View Article Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

2
Total Citations
View Citations
199
Total Downloads

Downloads (Last 12 months)20
Downloads (Last 6 weeks)2

Reflects downloads up to 18 Nov 2024

Other Metrics

View Author Metrics

Citations

Cited By

Kunze TTemplier PWilson D(2024)Searching Search Spaces: Meta-evolving a Geometric Encoding for Neural Networks2024 IEEE Congress on Evolutionary Computation (CEC)10.1109/CEC60901.2024.10612026(1-8)Online publication date: 30-Jun-2024
https://doi.org/10.1109/CEC60901.2024.10612026
Baratchi MWang CLimmer Svan Rijn JHoos HBäck TOlhofer M(2024)Automated machine learning: past, present and futureArtificial Intelligence Review10.1007/s10462-024-10726-157:5Online publication date: 18-Apr-2024
https://doi.org/10.1007/s10462-024-10726-1
Papp LHaberl DEcsedi BSpielvogel CKrajnc DGrahovac MMoradi SDrexler W(2023)DEBI-NNNeural Networks10.1016/j.neunet.2023.08.026167:C(517-532)Online publication date: 1-Oct-2023
https://dl.acm.org/doi/10.1016/j.neunet.2023.08.026

View Options

Login options

Check if you have access through your login credentials or your institution to get full access on this article.

Full Access

Get this Publication

View options

PDF

View or Download as a PDF file.

eReader

View online with eReader.

Media

Figures

Other

Tables

View Table of Contents