Emerging modularity during the evolution of neural networks

• Autorzy: Praczyk Tomasz

Identyfikatory

DOI

10.2478/jaiscr-2023-0010

• Języki publikacji: EN

Abstrakty

EN

Modularity is a feature of most small, medium and large–scale living organisms that has evolved over many years of evolution. A lot of artificial systems are also modular, however, in this case, the modularity is the most frequently a consequence of a handmade design process. Modular systems that emerge automatically, as a result of a learning process, are very rare. What is more, we do not know mechanisms which result in modularity. The main goal of the paper is to continue the work of other researchers on the origins of modularity, which is a form of optimal organization of matter, and the mechanisms that led to the spontaneous formation of modular living forms in the process of evolution in response to limited resources and environmental variability. The paper focuses on artificial neural networks and proposes a number of mechanisms operating at the genetic level, both those borrowed from the natural world and those designed by hand, the use of which may lead to network modularity and hopefully to an increase in their effectiveness. In addition, the influence of external factors on the shape of the networks, such as the variability of tasks and the conditions in which these tasks are performed, is also analyzed. The analysis is performed using the Hill Climb Assembler Encoding constructive neuro-evolutionary algorithm. The algorithm was extended with various module-oriented mechanisms and tested under various conditions. The aim of the tests was to investigate how individual mechanisms involved in the evolutionary process and factors external to this process affect modularity and efficiency of neural networks.

Słowa kluczowe

EN

neuro-evolution; modular neural networks; emergent modularity; hybrid algorithms

• Wydawca: University of Social Sciences
• Czasopismo: Journal of Artificial Intelligence and Soft Computing Research
• Rocznik: 2023
• Tom: Vol. 13, No. 2
• Strony: 107--126
• Opis fizyczny: Bibliogr. 50 poz., rys.

Twórcy

autor

Praczyk Tomasz

• Computer Department, Polish Naval Academy, ul. Smidowicza 69, 81-127 Gdynia, Poland

• https://orcid.org/0000-0003-0547-7935

Bibliografia

• [1] S. Ahmadian, S. Jalali, S. Islam, A. Khosravi, E. Fazli, and S. Nahavandi. A novel deep neuroevolution-based image classification method to diagnose coronavirus disease (covid19). Comput Biol Med., (139:104994), 2021.
• [2] A. Baldominos, Y. Saez, and P. Isasi. Evolutionary convolutional neural networks: an application to handwriting recognition. Neurocomputing, 283:38–52, 2018.
• [3] C.Y. Baldwin and K.B. Clark. Design Rules: The power of modularity. Chapter 3: What Is Modularity? MIT Press, 2018.
• [4] A. Billard and M. J. Mataric. Learning human movements by imitation: evaluation of a biologically inspired connectionist architecture. Robotics and Autonomous Systems, 941:1–16, 2001.
• [5] V. D. Blondel, J.-L. Guillaume, R. Lambiotte, and E. Lefebvre. Fast unfolding of communities in large networks. Journal of Statistical Mechanics Theory and Experiment, 10:10008, 2008.
• [6] R. Calabretta and J. Neirotti. Adaptive agents in changing environments, the role of modularity.Neural Process Lett, 42:257–274, 2015.
• [7] M. Carcenac. A modular neural network applied to image transformation and mental images. Neural Computing and Applications, 17:549–568, 2008.
• [8] J. Clune, B.E. Beckmann, P.K. McKinley, and C. Ofria. Investigating whether hyperneat produces modular neural networks. In Proceedings of the 12th annual conference on Genetic and evolutionary computation, pages 635–642, 2010.
• [9] J. Clune, J-B. Mouret, and H. Lipson. The evolutionary origins of modularity. In Proceedings of the Royal Society B, 2013.
• [10] [10] A.S. Cofino, J.M. Gutierrez, and M.L. Ivanissevich. Evolving modular networks with genetic algorithms: application to nonlinear time series. Expert Systems, 21(4):208–216, 2004.
• [11] Y. J. Cruz, M. Rivas, R. Quiza, A. Villalonga, R. E. Haber, and G. Beruvides. Ensemble of convolutional neural networks based on an evolutionary algorithm applied to an industrial welding process. Computers in Industry, 133:103530, 2021.
• [12] K. Deb, A. Pratap, S. Agarwal, , and T. Meyarivan. A fast and elitist multiobjective genetic algorithm: NSGA–II. IEEE Transactions on Evolutionary Computation, 6(2):182–197, 2002.
• [13] S. Doncieux and J. Meyer. Evolving modular neural networks to solve challenging control problems. In Proceedings of the Fourth International ICSC Symposium on Engineering of Intelligent Systems, 2004.
• [14] K. O. Ellefsen and J. Torresen. Evolving neural networks with multiple internal models. In Proceedings of the 14th European Conference on Artificial Life ECAL 2017, volume 14, pages 138–145, 2017.
• [15] K.O. Ellefsen, J-B. Mouret, and J. Clune. Neural modularity helps organisms evolve to learn new skills without forgetting old skills. PLoS Computational Biology, 11(4):e1004128, 2015.
• [16] C. Espinosa-Soto and A. Wagner. Specialization can drive the evolution of modularity. PLoS Computational Biology, 6(3):e1000719, 2010.
• [17] C. Fernando, D. Banarse, M. Reynolds, F. Besse, D. Pfau, M. Jaderberg, M. Lanctot, and D.Wierstra. Convolution by evolution: differentiable pattern producing networks. In Proceedings of the 2016 Genetic and Evolutionary Computation Conference, pages 109–116, 2016.
• [18] D. Filan, S. Hod, C. Wild, A. Critch, and S. Russell. Pruned neural networks are surprisingly modular. Technical Report arXiv:2003.04881[cs.NE], ArXiV, 2020.
• [19] J. Gauci and K. Stanley. Generating large–scale neural networks through discovering geometric regularities. In Proceedings of the Genetic and Evolutionary Computation Conference, pages 997–1004, 2007.
• [20] S. Han and S. Oh. An optimized modular neural network controller based on environment classification and selective sensor usage for mobile robot reactive navigation. Neural Computation and Application, 17:161–173, 2008.
• [21] J. Huizinga, J.B. Mouret, and J. Clune. Evolving neural networks that are both modular and regular: Hyperneat plus the connection cost technique. In Proceedings of the 2014 Annual Conference on Genetic and Evolutionary Computation, pages 697–704, 2014.
• [22] M. Hulse, S. Wischmann, and F. Pasemann. Structure and function of evolved neuro– controllers for autonomous robots. Connection Science, 16(4):249–266, 2004.
• [23] L. Kirsch, J. Kunze, and David Barber. Modular networks: Learning to decompose neural computation. Technical Report arXiv:1811.05249 [cs.LG], ArXiV, 2018.
• [24] J. Koutnik, J. Schmidhuber, and F. Gomez. Evolving deep unsupervised convolutional networks for vision–based reinforcement learning. In Proceedings of the 2014 Annual Conference on Genetic and Evolutionary Computation, pages 541–548, 2014.
• [25] V. Landassuri-Moreno and J. A. Bullinaria. Biasing the evolution of modular neural networks. In; 2011 IEEE Congress of Evolutionary Computation, 2011.
• [26] J. Liang, E. Meyerson, and R. Miikkulainen. Evolutionary architecture search for deep multitask networks. In GECCO ’18: Proceedings of the Genetic and Evolutionary Computation Conference, pages 466–473, 2018.
• [27] I. Loshchilov and F. Hutter. CMA–ES for hyperparameter optimization of deep neural networks. Technical Report arXiv: abs/1604.07269 [cs.NE],ArXiV, 2016.
• [28] P. Melin, D. Bravo, and O. Castillo. Fingerprint recognition using the fuzzy sugeno integral for response integration in modular neural networks. International Journal of General Systems, 37(4):499–515, 2008.
• [29] R. Miikkulainen, J. Liang, E. Meyerson, A. Rawal, D. Fink, O. Francon, B. Raju, H.; Shahrzad, A. Navruzyan, N. Duffy, and B. Hodjat. Evolving deep neural networks. Technical Report arXiv abs/1703.00548 [cs.NE], ArXiV, 2017.
• [30] J-B. Mouret and S. Doncieux. Evolving modular neural networks through exaptation. In 2009; IEEE Congress on Evolutionary Computation, ;pages 1570–1577, 2009.
• [31] H. Munn and M. Gallagher. Modularity in NEAT reinforcement learning networks, 2022.
• [32] N. NourAshrafoddin, A. R. Vahdat, and M. M.; Ebadzadeh. Automatic design of modular neural networks using genetic programming. In Proceedings of the 17th International Conference on Artificial Neural Networks ICANN 2007 Part I, pages 788–798, 2007.
• [33] M. Potter. The Design and Analysis of a Computational Model of Cooperative Coevolution. PhD; thesis, George Mason University, 1997.
• [34] M. A. Potter and K. A. De Jong. Cooperative coevolution: An architecture for evolving coadapted subcomponents. Evolutionary Computation, 8(1):1–29, 2000.
• [35] T. Praczyk. Probabilistic neural network application to warship radio stations identification. Computational Methods in Science and Technology, 13(1):53–58, 2007.
• [36] T. Praczyk. Using augmenting modular neural networks to evolve neuro–controllers for a team of underwater vehicles. Soft Computing, 18(12):2445–2460, 2014.
• [37] T. Praczyk. Cooperative co–evolutionary neural networks. Journal of Intelligent & Fuzzy Systems, 30(5):2843–2858, 2016.
• [38] T. Praczyk. Hill climb modular assembler encoding: Evolving modular neural networks of fixed modular architecture. Knowledge-Based Systems, 232:107493, nov 2021.
• [39] K. Soltanian, A. Ebnenasir, and M. Afsharchi. ;Modular grammatical evolution for the generation; of artificial neural networks. Evolutionary Computation, 30(2):291–327, 06 2022.
• [40] S. Sotirov, E. Sotirova, V. Atanassova, K. Atanassov, O. Castillo, P. Melin, T. Petkov, and S. Surchev. A hybrid approach for modular neural network design using intercriteria analysis and intuitionistic fuzzy logic. Complexity, 1:1–11, 2018.
• [41] K. O. Stanley and R. Miikkulainen. Evolving neural networks through augmenting topologies.; Evolutionary Computation, 10:99–127, 2002.
• [42] Y. Sun, B. Xue, M. Zhang, and G. G. Yen. Automatically designing CNN architectures using genetic algorithm for image classification. Technical; Report arXiv:1808.03818 [cs.NE], ArXiV, 2018.
• [43] C. R. Tosh. Can computational efficiency alone; drive the evolution of modularity in neural networks? Scientific Reports, 6:31982, 2016.
• [44] C. R. Tosh and L. McNally. The relative efficiency of modular and non–modular networks of different size. In Proceedings of the Royal Society B: Biological Sciences, volume 282:20142568, 2015.
• [45] A. Turan, S. D. Hinchberger, and M. H. El Naggar. Predicting the dynamic properties of glyben using a modular neural network (MNN). Canadian Geotechnical Journal, 45:1629–1638, 2008.
• [46] V. K. Valsalam and R. Miikkulainen. Evolving; symmetric and modular neural networks for distributed control. In Proceedings of the Genetic;and Evolutionary Computation Conference, 2009.
• [47] L. Xie and A. Yuille. Genetic CNN. Technical Report arXiv abs/1703.01513 [cs.NE], ArXiV, 2017.
• [48] X. Yao and Y. Liu. A new evolutionary system for evolving artificial neural networks. IEEE Transactions on Neural Networks, 8(3):694– 713,; 1997.
• [49] S.R. Young, D.C. Rose, T.P. Karnowsky, S.H. Lim, and R.M. Patton. Optimizing deep learning; hyper–parameters through an evolutionary algorithm. In Proceedings of the Workshop on Machine Learning in High–Performance Computing Environments, number 4, pages 1–5, 2015.
• [50] Z. Zhu, S. Guo, and M. Liao. Deep neuroevolution: Evolving neural network for character locomotion controller. In 2021 2nd International Conference on Artificial Intelligence and Information Systems, ICAIIS 2021, New York, NY, USA,; 2021. Association for Computing Machinery.

Uwagi

Opracowanie rekordu ze środków MEiN, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2022-2023).