Recurrent neural networks have the potential to perform significantly better than the commonly used feedforward neural networks due to their dynamical nature. However, they have received less attention because training algorithms/architectures have not been well developed. In this study, a recursive least squares algorithm to train recurrent neural networks with an arbitrary number of hidden layers is developed. The training algorithm is developed as an extension of the standard recursive estimation problem. Simulated results obtained for identification of the dynamics of a nonlinear dynamical system show promising results.
Q. Xu et al., "A Recursive Least Squares Training Algorithm for Multilayer Recurrent Neural Networks," Proceedings of the American Control Conference (1994, Baltimore, MD), vol. 2, pp. 1712-1716, Institute of Electrical and Electronics Engineers (IEEE), Jun 1994.
The definitive version is available at https://doi.org/10.1109/ACC.1994.752364
American Control Conference (1994: Jun. 29-Jul. 1, Baltimore, MD)
Mechanical and Aerospace Engineering
Keywords and Phrases
Feedforward Neural Networks; Learning (Artificial Intelligence); Least Squares Approximations; Multilayer Perceptrons; Multilayer Recurrent Neural Networks; Nonlinear Dynamical System; Recurrent Neural Nets; Recursive Estimation; Recursive Least Squares Training Algorithm; Training Architectures; Convergence of numerical methods; Identification (control systems); Learning systems; Least squares approximations; Neural networks; Nonlinear control systems; Recursive functions; Nonlinear dynamical systems; Algorithms
International Standard Serial Number (ISSN)
Article - Conference proceedings
© 1994 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.