Improved broad learning system: partial weights modification based on BP algorithm

2018 International Conference on Advanced Electronic Materials, Computers and Materials Engineering

材料科学;无线电电子学;计算机科学

Li, Da^1 ; Shujuan, Ji^2 ; Chunjin, Zhang^3

College of Computer Science and Technology, Shandong University of Science and Technology, Qingdao, Shandong

266590, China^1

Key Laboratory for Wisdom Mine Information Technology of Shandong Province, Shandong University of Science and Technology, Qingdao, Shandong

266590, China^2

Network Information Center (NIC), Shandong University of Science and Technology, Qingdao, Shandong

266590, China^3

关键词: BP algorithm;    Feature nodes;    Random generation;    Single layer feed-forward neural networks;    Training process;    Training time;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/439/3/032083/pdf
DOI  :  10.1088/1757-899X/439/3/032083

来源: IOP

【 摘 要 】

Although breakthrough achievements of deep learning have been made in different areas, there is no good idea to prevent the time-consuming training process. Single-layer feedforward neural networks (e.g. BLS) are used to reduce the training time. However, with the decrease of training time, the accuracy degradation has emerged. In view of the limitation of random generation of connection parameters between feature nodes and enhancement nodes, this paper presents an algorithm (IBLS) based on BLS and backpropagation algorithm to learn the weights between feature nodes and enhancement nodes. Experiments over NORB and MNIST data sets show that the improved broad learning system achieves acceptable results.