Regular PaperNo Access

Evolving Deep Neural Networks with Cultural Algorithms for Real-Time Industrial Applications

Faisal Waris

Department of Computer Science, Wayne State University, Detroit, MI 48201, USA

E-mail Address: fwaris@wayne.edu

Search for more papers by this author

Robert G. Reynolds

Department of Computer Science, Wayne State University, Detroit, MI 48201, USA

Museum of Anthropological Archaeology, University of Michigan, -Ann Arbor, MI 48104, USA

E-mail Address: reynolds@cs.wayen.edu

Search for more papers by this author

, and

Joonho Lee

Guidance Navigation Control & Autonomy, Boeing Research and Technology, Everett, Washington 98204, USA

E-mail Address: joonho.lee5@boeing.com

Search for more papers by this author

https://doi.org/10.1142/S1793351X22500027Cited by:4 (Source: Crossref)

Abstract

The goal of this paper is to investigate the applicability of evolutionary algorithms to the design of real-time industrial controllers. Present-day “deep learning” (DL) is firmly established as a useful tool for addressing many practical problems. This has spurred the development of neural architecture search (NAS) methods in order to automate the model search activity. CATNeuro is a NAS algorithm based on the graph evolution concept devised by Neuroevolution of Augmenting Topologies (NEAT) but propelled by cultural algorithm (CA) as the evolutionary driver. The CA is a network-based, stochastic optimization framework inspired by problem solving in human cultures. Knowledge distribution (KD) across the network of graph models is a key to problem solving success in CAT systems. Two alternative mechanisms for KD across the network are employed. One supports cooperation (CATNeuro) in the network and the other competition (WM). To test the viability of each configuration prior to use in the industrial setting, they were applied to the design of a real-time controller for a two-dimensional fighting game. While both were able to beat the AI program that came with the fighting game, the cooperative method performed statistically better. As a result, it was used to track the motion of a trailer (in lateral and vertical directions) using a camera mounted on the tractor vehicle towing the trailer. In this second real-time application (trailer motion), the CATNeuro configuration was compared to the original NEAT (elitist) method of evolution. CATNeuro is found to perform statistically better than NEAT in many aspects of the design including model training loss, model parameter size, and overall model structure consistency. In both scenarios, the performance improvements were attributed to the increased model diversity due to the interaction of CA knowledge sources both cooperatively and competitively.

Keywords:

References

1. I. Goodfellow, Y. Bengio and A. Courville , Deep Learning (MIT Press, Cambridge, MA, 2016). Google Scholar
2. K. He, X. Zhang, S. Ren and J. Sun, Deep residual learning for image recognition, arXiv:1512.03385, 2015. Google Scholar
3. J. Devlin, M.-W. Chang, K. Lee and K. Toutanova , BERT: Pre-training of deep bidirectional transformers for language understanding, in NAACL-HLT 2019: Annual Conf. North American Chapter of the Association for Computational Linguistics, 2019, pp. 4171–4186. Google Scholar
4. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser and I. Polosukhin , Attention is all you need, in Proc. 31st Int. Conf. Neural Information Processing Systems, 2017, pp. 6000–60101. Google Scholar
5. H. Liu, K. Simonyan and Y. Yang , DARTS: Differentiable architecture search, in ICLR 2019: 7th Int. Conf. Learning Representations, 2019, pp. 6000–6010. Google Scholar
6. Z. Guo, X. Zhang, H. Mu, W. Heng, Z. Liu, Y. Wei and J. Sun, Single path one-shot neural architecture search with uniform sampling, arXiv:1904.00420, 2019. Google Scholar
7. E. Bisong , Google AutoML: Cloud Vision, Bilding Machine Learning and Deep Learning Models on Google Cloud Platform. Aperss, Berkeley, CA. https://doi.org/10.1007/978-1-4842-4470-8_42. Google Scholar
8. Microsoft, Neural Network Intelligence, 2020, https://github.com/microsoft/nni. Google Scholar
9. K. Stanley and R. Miikkulainen , Efficient evolution of neural network topologies, in Proc. 2002 Congress on Evolutionary Computation. CEC’02 (Cat. No. 02TH8600), 2002, pp. 1757–1762. Crossref, Google Scholar
010. R. Miikkulainen, J. Z. Liang, E. Meyerson, A. Rawal, D. Fink, O. Francon, B. Raju, H. Shahrzad, A. Navruzyan, N. Duffy and B. Hodjat, Evolving deep neural networks, arXiv:1703.00548. Google Scholar
11. M. Z. Ali, P. N. Suganthan, R. G. Reynolds and A. F. Al-Badarneh , Leveraged neighborhood restructuring in cultural algorithms for solving real-world numerical optimization problems, IEEE Trans. Evol. Comput. 20(2) [2016] 218–231. Crossref, Web of Science, Google Scholar
12. R. Reynolds, M. Ali and T. Jayyousi , Mining the social fabric of archaic urban centers with cultural algorithms, IEEE Comput. 41(1) [2008] 64–72. Crossref, Web of Science, Google Scholar
13. T. Judeh, T. Jayyousi, L. R. Acharya, R. G. Reynolds and D. Zhu , GSCA: Reconstructing biological pathway topologies using a cultural algorithms approach, in 2014 IEEE Congress on Evolutionary Computation, 2014, pp. 2206–2213. Crossref, Google Scholar
14. F. Waris and R. G. Reynolds , Human-defined, machine-optimized — Gesture recognition using a mixed approach, in 2015 Swarm/Human Blended Intelligence Workshop, 2015, pp. 1–9. Crossref, Google Scholar
15. F. Waris and R. G. Reynolds , Optimizing AI pipelines: A game-theoretic cultural algorithms approach, in 2018 IEEE Congress on Evolutionary Computation, 2018, pp. 1–10. Crossref, Google Scholar
16. C. Shalizi, Estimating distributions and densities, 2009, http://www.stat.cmu.edu/ $\sim$ cshalizi/350/lectures/28/lecture-28.pdf. Google Scholar
17. Y. Shi , Brain storm optimization algorithm, in ICSI’11 Proc. Second Int. Conf. Advances in Swarm Intelligence — Volume Part I, 2011, pp. 303–309. Crossref, Google Scholar
18. A. Al-Tirawi and R. G. Reynolds , Using common value auction in cultural algorithm to enhance robustness and resilience of social knowledge distribution systems, in 2018 IEEE Symp. Series on Computational Intelligence, 2018, pp. 755–764. Crossref, Google Scholar
19. R. Reynolds and L. Kinnaird-Heether , Optimization problem solving with auctions in cultural algorithms, Memetic Comput. 5(2) [2013] 83–94. Crossref, Web of Science, Google Scholar
20. J. W. Boudreau, L. Rentschler and S. Sanders , Stag-hunt contests and alliance formation, Public Choice 179(3–4) [2019] 267. Crossref, Web of Science, Google Scholar
21. B. Skyrms , The Stag Hunt and Evolution of Social Structure (Cambridge University Press, Cambridge, UK, 2004). Google Scholar
22. Intelligent Computer Entertainment Lab., Ritsumeikan University, Fighting game AI competition, 2018, http://www.ice.ci.ritsumei.ac.jp/ $\sim$ ftgaic/. Google Scholar
23. R. S. Sutton, D. A. McAllester, S. P. Singh and Y. Mansour , Policy gradient methods for reinforcement learning with function approximation, in Advances in Neural Information Processing Systems 12, 1999, pp. 1057–1063. Google Scholar
24. J. Kober, J. A. Bagnell and J. Peters , Reinforcement learning in robotics: A survey, Int. J. Robot. Res. 32(11) [2013] 1238–1274. Crossref, Web of Science, Google Scholar
25. C. Y. Chu and R. Thawonmas , Applying fuzzy control in fighting game AI, in Proc. 77th Convention of Information Processing Society of Japan, 2017, pp. 99–100. Google Scholar
26. S. T. Roweis and L. K. Saul , Nonlinear dimensionality reduction by locally linear embedding, Science 290(5500) [2000] 2323–2326. Crossref, Web of Science, Google Scholar
27. F. Seide and A. Agarwal , CNTK: Microsoft’s open-source deep-learning toolkit, in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, 2016, p. 2135. Crossref, Google Scholar
28. D. B. Post, AlphaGo Zero: Starting from scratch, October 18, 2017, https://deepmind.com/blog/article/alphago-zero-starting-scratch [accessed 29 August 2020]. Google Scholar
29. Y. Bengio , Deep learning of representations: Looking forward, in SLSP’13 Proc. First Int. Conf. Statistical Language and Speech Processing, 2013, pp. 1–37. Crossref, Google Scholar
30. T. Shim and S. Chen , Investigation of trailer yaw motion control using active front steer and differential brake, SAE Int. 4(1) [2011] 1057–1107. Google Scholar
31. S. K. Jukka Ahokas , Dynamic behavior of a tractor–trailer combination during braking, Biosyst. Eng. 85(1) [2003] 29–39. Crossref, Web of Science, Google Scholar
32. C. Harris and M. Stephens , A combined corner and edge detector, in Proc. Fourth Alvey Vision Conf., 1988, pp. 23.1–23.6. Crossref, Google Scholar
33. B. D. Lucas and T. Kanade , An iterative image registration technique with an application to stereo vision, in Proc. Imaging Understanding Workshop, 1981, pp. 121–130. Google Scholar