Abstract
The learning algorithms of multilayered feed-forward networks can be classified into two categories, gradient and non-gradient kinds. The gradient descent algorithms like backpropagation (BP) or its variations are widely used in many application areas because of convenience. However, the most serious problem associated with the BP is local minima problem. We propose an improved gradient descent algorithm intended to weaken the local minima problem without doing any harm to simplicity of the gradient descent method. This algorithm is called dual gradient learning algorithm in which the upper connections (hidden-to-output) and the lower connections (input-to-hidden) separately evaluated and trained. To do so, the target values of hidden layer units are introduced to be used as evaluation criteria of the lower connections. Simulations on some benchmark problems and a real classification task have been performed to demonstrate the validity of the proposed method.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Cetin 1, B., Barhen, J.: Global descent replaces gradient descent to avoid local minima problem in learning with ANN. In: Proc. of lEEE Conf. on NN, vol. 2, pp. 836–842 (1993)
Hont, R., Padalos, P. (eds.): Handbook of Global Optimization. Kluwer, Dordrecht (1995)
Plagianakos, P., Magoulas, G.D., Vrahatis, M.: Deterministic non-monotone strategies for the effecive training of multilayer perceptrons. IEEE Trans. on N. Networks 13(6), 1268–1284 (2002)
Jones, D., Perttunen, C., Stuckman, B.: Lipschitzian Optimization without the Lipschitz Constant. J. of Optimization Theory and Applications 79, 157–181 (1993)
Vrahatis, M., Androulakis, G., Lambrinos, J., Magoulas, G.: A class of gradient unconstrained minimization stepsize. J. of Computational and Applied Mathematics 114, 367–386 (2000)
Jordanov, I.N., Rafik, T.A.: Local Minima Free Network Learning. In: Second lEEE International Conference On Intelligent Systems, pp. 34–39 (June 2004)
Bilbro, G.: Fast stochastic global optimisation. IEEE Trans. On Systems, Man, and Cybernetics 24, 684–689 (1994)
Huyer, W., Neumaier, A.: Global Optimization by Multilevel Coordinate Search. J. of Gl. Optimizalion 14, 331–335 (1999)
Tom, A., Vitanen, S.: Topographical Global Optimization Using Pre-Sampled Points. J. of Global Optimization 5, 267–276 (1994)
Ng, S.C., Leung, S.H., Luk, A.: A Hybrid Algorithm of Weight Evolution and Generalized Back-propagation for finding Global Minimum. In: Proceedings of IEEE International Joint Conference on Neural Networks (IJCNN’99), IEEE Computer Society Press, Los Alamitos (1999)
Yu, X.H.: Can backpropagation error surface not have local minima. IEEE Transactions on Neural Networks 3(6), 1019–1021 (1992)
Yu, X.H., Chen, G.A.: On the local minima free condition of backpropagation learning. IEEE Transactions on Neural Networks 6(5), 1300–1303 (1995)
Wang, C., Principe, J.C.: Training neural networks with additive noise in the desired signal. IEEE Trans. Neural Networks 10(6), 1511–1517 (1999)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 2007 Springer Berlin Heidelberg
About this paper
Cite this paper
Choi, B., Lee, JH., Park, TS. (2007). Dual Gradient Descent Algorithm on Two-Layered Feed-Forward Artificial Neural Networks. In: Okuno, H.G., Ali, M. (eds) New Trends in Applied Artificial Intelligence. IEA/AIE 2007. Lecture Notes in Computer Science(), vol 4570. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-73325-6_69
Download citation
DOI: https://doi.org/10.1007/978-3-540-73325-6_69
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-73322-5
Online ISBN: 978-3-540-73325-6
eBook Packages: Computer ScienceComputer Science (R0)