Meta-Learning for Adaptive Identification of Non-Linear Dynamical Systems
- 1 January 2005
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21589860,p. 473-478
- https://doi.org/10.1109/.2005.1467061
Abstract
Adaptive identification of non-linear dynamical systems via recurrent neural networks (RNNs) is presented in this paper. We explore the notion that a fixed-weight RNN needs to change only its internal state to change its behavior policy. This ability is acquired through prior training procedure that enable the learning of adaptive behaviors. Some simulation results are presentedKeywords
This publication has 12 references indexed in Scilit:
- Identification and Control of a Nonlinear Discrete-Time System Based on its Linearization: A Unified FrameworkIEEE Transactions on Neural Networks, 2004
- Adaptation from fixed weight dynamic networksPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Fixed-weight controller for multiple systemsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- A recurrent fuzzy-neural model for dynamic system identificationIEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2002
- New results on recurrent network training: unifying the algorithms and accelerating convergenceIEEE Transactions on Neural Networks, 2000
- Fixed-weight on-line learningIEEE Transactions on Neural Networks, 1999
- Long Short-Term MemoryNeural Computation, 1997
- Adaptive behavior from fixed weight networksInformation Sciences, 1997
- Application of the recurrent multilayer perceptron in modeling complex process dynamicsIEEE Transactions on Neural Networks, 1994
- Identification and control of dynamical systems using neural networksIEEE Transactions on Neural Networks, 1990