Quantized Neural Modeling: Hybrid Quantized Architecture in Elman Networks
- 310 Downloads
This paper presents a novel neural network model with hybrid quantized architecture to improve the performance of the conventional Elman networks. The quantum gate technique is introduced for solving the pattern mismatch between the inputs stream and one-time-delay state feedback. A quantized back-propagation training algorithm with an adaptive dead zone scheme is developed for providing an optimal or suboptimal tradeoff between the convergence speed and the generalization performance. Furthermore, the effectiveness of the new real time learning algorithm is demonstrated by proving the quantum gate parameter convergence based on Lyapunov method. The numerical experiments are carried out to demonstrate the accuracy of the theoretical results.
KeywordsElman networks Hybrid quantized structure Quantum Elman back-propagation training Quantized parameter convergence Generalization performance
Unable to display preview. Download preview PDF.
- 2.Jordan M (1986) Serial order: a parallel distributed processing approach. ICS Report 8604, Institute of Cognitive Science, University of California San Diego, San DiegoGoogle Scholar
- 15.Menneer T, Narayanan A (1995) Quantum-inspired neural networks. Technical Report R329, University of Exeter, ExeterGoogle Scholar
- 19.Matsui N, Takai M, Nishimura H (1988) A network model based on qubit-like neuron corresponding to quantum circuit. IEICE J81-A(12):1687–1692Google Scholar
- 21.Li PH, Chai Y, Xiong QY (2011) Quantized neuronal modeling: quantum gate structure in Elman networks. In: Proceedings of fourth international workshop on advanced computational intelligence. IEEE Press, Wu Han, pp 315–320Google Scholar
- 23.Vidyasagar M (1980) Nonlinear systems analysis. Prentice-Hall, Englewood CliffsGoogle Scholar