A broad learning system based on reservoir computing

被引：0
|
作者：
Yang G. [1 ,2 ]
Chen P. [1 ,2 ]
Dai L.-Z. [1 ,2 ]
Yang H. [1 ,2 ]
机构：
[1] School of Electrical and Automation Engineering, East China Jiaotong University, Nanchang
[2] Key Laboratory of Advanced Control & Optimization of Jiangxi Province, Nanchang
来源：
Kongzhi yu Juece/Control and Decision | 2021年 / 36卷 / 09期
关键词：
Broad learning; Echo state network; Incremental learning; Reservoir computing; Time series prediction;
D O I：
10.13195/j.kzyjc.2019.1729
中图分类号：
O212 [数理统计];
学科分类号：
摘要：
Broad learning system (BLS), which has characteristics of fast and accuracy, is an efficient incremental learning systems based on random vector function-link network (RVFLN). In order to realize the precise prediction of time-series, a broad learning system based on reservoir computing reservoir computing broad learning systems (RCBLS) is proposed combined with the reservoir structure of echo state network (ESN). A simple circle reservoir connection is introduced in the RCBLS's enhancement layer to replace the feedforward connection of BLS, which makes the RCBLS have certain echo state characteristics and convenient for design. At the same time, incremental learning is applied to ensure RCBLS's real-time performance. Based on the multiple superimposed oscillator (MSO) time series prediction problems, the performance of the RCBLS with different reservoir structures under different scales of data sample is studied respectively. The results show that the RCBLS with multi-reservoir structure improves the generalization performance and stability greatly. © 2021, Editorial Office of Control and Decision. All right reserved.
引用
收藏
页码：2203 / 2210
页数：7
相关论文
共 29 条
  • [1] Chen C L P, Liu Z., Broad learning system: An effective and efficient incremental learning system without the need for deep architecture, IEEE Transactions on Neural Networks and Learning Systems, 29, 1, pp. 10-24, (2018)
  • [2] Chen C L P, Wan J Z., A rapid learning and dynamic stepwise updating algorithm for flat neural networks and the application to time-series prediction, IEEE Transactions on Systems, Man, and Cybernetics: Part B, 29, 1, pp. 62-72, (1999)
  • [3] Chen C L P., A rapid supervised learning neural network for function interpolation and approximation, IEEE Transactions on Neural Networks, 7, 5, pp. 1220-1230, (1996)
  • [4] Igelnik B, Pao Y., Stochastic choice of basis functions in adaptive function approximation and the functional-link net, IEEE Transactions on Neural Networks, 6, 6, pp. 1320-1329, (1995)
  • [5] Pao Y H, Park G H, Sobajic D J., Learning and generalization characteristics of the random vector functional-link net, Neurocomputing, 6, 2, pp. 163-180, (1994)
  • [6] Pao Y H, Takefuji Y., Functional-link net computing: Theory, system architecture, and functionalities, Computer, 25, 5, pp. 76-79, (1992)
  • [7] Chen C, Liu Z, Feng S., Universal approximation capability of broad learning system and its structural variations, IEEE Transactions on Neural Networks Learning Systems, 30, 4, pp. 1191-1204, (2019)
  • [8] Jin J W, Chen C L P., Regularized robust broad learning system for uncertain data modeling, Neurocomputing, 322, pp. 58-69, (2018)
  • [9] Kong Y, Wang X, Cheng Y, Et al., Hyperspectral imagery classification based on semi-supervised broad learning system, Remote Sensing, 10, 5, (2018)
  • [10] Liu W, Yang H, Sun Y, Et al., A broad neural network structure for class incremental learning, International Symposium on Neural Networks, pp. 229-238, (2018)
123