Composite FORCE learning of chaotic echo state networks for time-series prediction
This work addresses the challenge of enhancing prediction accuracy for chaotic time series, which is important for fields like signal processing and forecasting, but it is incremental as it builds on existing FORCE learning techniques.
The paper tackled the problem of training echo state networks (ESNs) with chaotic initial activity for time-series prediction by proposing a composite FORCE learning method based on recursive least squares, which significantly improved learning and prediction performances on the Mackey-Glass chaotic benchmark compared to existing methods.
Echo state network (ESN), a kind of recurrent neural networks, consists of a fixed reservoir in which neurons are connected randomly and recursively and obtains the desired output only by training output connection weights. First-order reduced and controlled error (FORCE) learning is an online supervised training approach that can change the chaotic activity of ESNs into specified activity patterns. This paper proposes a composite FORCE learning method based on recursive least squares to train ESNs whose initial activity is spontaneously chaotic, where a composite learning technique featured by dynamic regressor extension and memory data exploitation is applied to enhance parameter convergence. The proposed method is applied to a benchmark problem about predicting chaotic time series generated by the Mackey-Glass system, and numerical results have shown that it significantly improves learning and prediction performances compared with existing methods.