We also indicate the locations along this spiral in the parameter-space, where such bifurcations happen.
📌 3 Key Takeaways
🎯
Objective & Hypothesis:In this work we report numerical results involving a certain Hopfield-type three-neurons network, with the hyperbolic tangent as the activation function.
📊
Key Findings & Data:Specifically, we investigate a place of a two-dimensional parameter-space of this system where typical periodic structures, the so-called shrimps, are embedded in a chaotic region.
💡
Conclusion & Impact:We also indicate the locations along this spiral in the parameter-space, where such bifurcations happen.
🔬Design:Research Article
📍Scope:Comprehensive Sample
⚙️Variables:Chaos · Lyapunov exponents spectrum · Parameter space
Abstract
In this work we report numerical results involving a certain Hopfield-type three-neurons network, with the hyperbolic tangent as the activation function. Specifically, we investigate a place of a two-dimensional parameter-space of this system where typical periodic structures, the so-called shrimps, are embedded in a chaotic region. We show that these structures are organized themselves as a spiral that coil up toward a focal point, while undergo period-adding bifurcations. We also indicate the locations along this spiral in the parameter-space, where such bifurcations happen.
SİLVA, A. D., RECH, P. (2022). On Bifurcations Along the Spiral Organization of the Periodicity in a Hopfield Neural Network. Communications in Advanced Mathematical Sciences, Cilt: 5 - Sayı: 2, 92–98. https://ajindex.com/articles/748219
J. J. Hopfield, Neurons with graded response have collective computational properties like those of two-state neurons, Proc. Natl. Acad. Sci. USA, 81 (1984), 3088–3092
E. Korner, R. Kupper, M. K. M. Rahman, Y. Shkuro, Neurocomputing Research Developments, Nova Science Publishers, New York, 2007
W. Z. Huang, Y. Huang, Chaos, bifurcations and robustness of a class of Hopfield neural networks, Int. J. Bifurcation and Chaos, 21 (2011), 885–895
P. F. Chen, Z. Q. Chen, and W. J. Wu, A novel chaotic system with one source and two saddle-foci in Hopfield neural networks, Chin. Phys. B, 19 (2010), 040509
P. Zheng, W. Tang, J. Hang, Some novel double-scroll chaotic attractors in Hopfield networks, Neurocomputing, 73 (2010), 2280–2285
A. C. Mathias and P. C. Rech, Hopfield neural network: The hyperbolic tangent and the piecewise-linear activation functions, Neural Networks, 34 (2012), 42–45
P. C. Rech, Period-adding and spiral organization of the periodicity in a Hopfield neural network, Int. J. Mach. Learn. & Cyber., 6 (2015), 1–6
A. Wolf, J. B. Swift, H. L. Swinney, and J. A. Vastano, Determining Lyapunov exponents from a time series, Physica D, 16 (1985), 285–317
C. Bonatto, J. A. C. Gallas, Periodicity hub and nested spirals in the phase diagram of a simple resistive circuit, Phys. Rev. Lett., 101 (2008), 054101
J. A. C. Gallas, The structure of infinite periodic and chaotic hub cascades in phase diagrams of simple autonomous flows, Int. J. Bifurcation and Chaos, 20 (2010), 197–211
H. A. Albuquerque, P. C. Rech, Spiral periodic structure inside chaotic region in parameter-space of a Chua circuit, Int. J. Circ. Theor. Appl., 40 (2012), 189–194
R. Barrio, F. Blesa, S. Serrano, Qualitative analysis of the R¨ossler equations: Bifurcations of limit cycles and chaotic attractors, Physica D, 238 (2009), 1087–1100
X. F. Li, Y. T. L. Andrew, Y. D. Chu, Symmetry and period-adding windows in a modified optical injection semiconductor laser model, Chin. Phys. Lett., 29 (2012), 010201
C. Stegemann, P. C. Rech, Organization of the dynamics in a parameter plane of a tumor growth mathematical model, Int. J. Bifurcation and Chaos, 24 (2014), 1450023
R. A. da Silva, P. C. Rech, Spiral periodic structures in a parameter plane of an ecological model, Appl. Math. Comput., 254 (2015), 9–13
P. C. Rech, Spiral organization of periodic structures in the Lorenz-Stenflo system, Phys. Scr., 91 (2016), 075