Abstract
The gradient method for training Elman networks with a finite training sample set is considered. Monotonicity of the error function in the iteration is shown. Weak and strong convergence results are proved, indicating that the gradient of the error function goes to zero and the weight sequence goes to a fixed point, respectively. A numerical example is given to support the theoretical findings.
Similar content being viewed by others
References
Elman J L. Finding structure in time[J]. Cognitive Science, 1990, 14(2):179–211.
Tsoi A C, Back A D. Locally recurrent globally feedforward networks: a critical review of architectures[J]. IEEE Transactions on Neural Networks, 1994, 5(2):229–239.
Wang Deliang, Liu Xiaomei, Ahalt S C. On temporal generalization of simple recurrent networks[J]. Neural Networks, 1996, 9(7):1099–1118.
Kremer S C. On the computational power of Elman-style recurrent networks[J]. IEEE Transactions on Neural Networks, 1995, 6(4):1000–1004.
Pham D T, Liu X. Training of Elman networks and dynamic system modeling[J]. International Journal of Systems Science, 1996, 27(2):221–226.
Cartling B. On the implicit acquisition of a context-free grammar by a simple recurrent neural network[J]. Neurocomputing, 2008, 71(7–9):1527–1537.
Li Xiang, Chen Zengqiang, Yuan Zhuzhi, et al. Generating chaos by an Elman network[J]. IEEE Transactions on Circuits and Systems-I, 2001, 48(9):1126–1131.
Ekici S, Yildirim S, Poyraz M. A transmission line fault locator based on Elman recurrent networks[J]. Applied Soft Computing, 2008, DOI: 10.1016/j.asoc.2008.04.011.
Neto L B, Coelho P H G, Soares de Mello J C C B, et al. Flow estimation using an Elman networks[C]. In: Wunsch D, et al (eds). Proceedings of 2004 IEEE International Joint Conference on Neural Networks, IEEE Press, Budapest, Hungary, 2004, 831–836.
Demuth H B, Beale M H, Hagan M T. Neural network toolbox user’s guide[M]. Natick, MA: The Mathworks Inc, 2007.
Jesús O D, Hagan M T. Backpropation algorithms for a broad class of dynamic networks[J]. IEEE Transactions on Neural Networks, 2007, 18(1):14–27.
Williams R J, Zisper D. A learning algorithm for continually running fully recurrent neural networks[J]. Neural Computation, 1989, 1(2):270–280.
Ku C C, Lee K Y. Diagonal recurrent neural networks for dynamic systems control[J]. IEEE Transaction on Neural Networks, 1995, 6(1):144–156.
Xu Dongpo, Li Zhengxue, Wu Wei, et al. Convergence of gradient descent algorithm for diagonal recurrent neural networks[C]. In: Cui Guangzhao, et al (eds). International Conference on Bio-Inspired Computing: Theories and Applications, IEEE Press, Zhengzhou, China, 2007.
Kuan C M, Hornik K, White H. A convergence results for learning in recurrent neural networks[J]. Neural Computation, 1994, 6(3):420–440.
Wu Wei, Feng Guorui, Li Zhengxue, et al. Convergence of an online gradient method for BP neural networks[J]. IEEE Transactions on Neural Networks, 2005, 16(3):533–540.
Wu Wei, Shao Hongmei, Qu Di. Strong convergence for gradient methods for BP networks training[C]. In: Zhao Mingsheng and Shi Zhongzhi (eds). Proceedings of 2005 International Conference on Neural Networks and Brains, IEEE Press, Beijing, China, 2005, 332–334.
Gori M, Maggini M. Optimal convergence of on-line backpropagation[J]. IEEE Transaction on Neural Networks, 1996, 7(1):251–254.
Ortega J, Rheinboldt W. Iterative solution of nonlinear equations in several variables[M]. New York: Academic Press, 1970.
Yuan Yaxiang, Sun Wenyu. Optimization theory and methods[M]. Beijing: Science Press, 2001, 149 (in Chinese).
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by GUO Xing-ming
Project supported by the National Natural Science Foundation of China (No. 10471017)
Rights and permissions
About this article
Cite this article
Wu, W., Xu, Dp. & Li, Zx. Convergence of gradient method for Elman networks. Appl. Math. Mech.-Engl. Ed. 29, 1231–1238 (2008). https://doi.org/10.1007/s10483-008-0912-z
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10483-008-0912-z