RCBLS: An Outlier-Robust Broad Learning Framework with Compact Structure

被引:2
作者
Guo, Wei [1 ,2 ]
Yu, Jianjiang [2 ]
Zhou, Caigen [2 ]
Yuan, Xiaofeng [2 ]
Wang, Zhanxiu [2 ]
机构
[1] Yancheng Teachers Univ, Jiangsu Prov Key Constructive Lab Big Data Psychol, Yancheng 224002, Peoples R China
[2] Yancheng Teachers Univ, Coll Informat Engn, Yancheng 224002, Peoples R China
基金
中国国家自然科学基金;
关键词
broad learning system; robustness; compactness; M-estimator; sparsity regularization; SYSTEM; APPROXIMATION; CAPABILITY; FEATURES;
D O I
10.3390/electronics12143118
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Recently, the broad learning system (BLS) has been widely developed in virtue of its excellent performance and high-computational efficiency. However, two deficiencies still exist in BLS and preclude its deployment in real applications. First, the standard BLS performs poorly in outlier environments because the least squares loss function it uses to train the network is sensitive to outliers. Second, the model structure of BLS is likely to be redundant since the hidden nodes in it are randomly generated. To address the above two issues, a new robust and compact BLS (RCBLS), based on M-estimator and sparsity regularization, is proposed in this paper. The RCBLS develops from the BLS model and maintains its excellent characteristics, but replaces the conventional least squares learning criterion with an M-estimator-based loss function that is less sensitive to outliers, in order to suppress the incorrect feedback of the model to outlier samples, and hence enhance its robustness in the presence of outliers. Meanwhile, the RCBLS imposes the sparsity-promoting l2,1 -norm regularization instead of the common l2-norm regularization for model reduction. With the help of the row sparsity of l2,1-norm regularization, the unnecessary hidden nodes in RCBLS can be effectively picked out and removed from the network, thereby resulting in a more compact network. The theoretical analyses on outlier robustness, structural compactness and computational complexity of the proposed RCBLS model are provided. Finally, the validity of the RCBLS is verified by regression, time series prediction and image classification tasks. The experimental results demonstrate that the proposed RCBLS has stronger anti-outlier ability and more compact network structure than BLS and other representative algorithms.
引用
收藏
页数:23
相关论文
共 60 条
[1]   Broad Learning Based Dynamic Fuzzy Inference System With Adaptive Structure and Interpretable Fuzzy Rules [J].
Bai, Kaiyuan ;
Zhu, Xiaomin ;
Wen, Shiping ;
Zhang, Runtong ;
Zhang, Wenyu .
IEEE TRANSACTIONS ON FUZZY SYSTEMS, 2022, 30 (08) :3270-3283
[2]   Incremental Learning for Remaining Useful Life Prediction via Temporal Cascade Broad Learning System With Newly Acquired Data [J].
Cao, Yudong ;
Jia, Minping ;
Ding, Peng ;
Zhao, Xiaoli ;
Ding, Yifei .
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, 2023, 19 (04) :6234-6245
[3]  
Chang Victor, 2023, International Journal of Business Information Systems, P503, DOI 10.1504/IJBIS.2023.129698
[4]   Presenting Cloud Business Performance for Manufacturing Organizations [J].
Chang, Victor .
INFORMATION SYSTEMS FRONTIERS, 2020, 22 (01) :59-75
[5]   Random-Positioned License Plate Recognition Using Hybrid Broad Learning System and Convolutional Networks [J].
Chen, C. L. Philip ;
Wang, Bingshu .
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 2022, 23 (01) :444-456
[6]   Universal Approximation Capability of Broad Learning System and Its Structural Variations [J].
Chen, C. L. Philip ;
Liu, Zhulin ;
Feng, Shuang .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2019, 30 (04) :1191-1204
[7]   Broad Learning System: An Effective and Efficient Incremental Learning System Without the Need for Deep Architecture [J].
Chen, C. L. Philip ;
Liu, Zhulin .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2018, 29 (01) :10-24
[8]   Frequency Principle in Broad Learning System [J].
Chen, Guang-Yong ;
Gan, Min ;
Chen, C. L. Philip ;
Zhu, Hong-Tao ;
Chen, Long .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2022, 33 (11) :6983-6989
[9]   Double-kernel based class-specific broad learning system for multiclass imbalance learning [J].
Chen, Wuxing ;
Yang, Kaixiang ;
Yu, Zhiwen ;
Zhang, Weiwen .
KNOWLEDGE-BASED SYSTEMS, 2022, 253
[10]   Weighted Broad Learning System and Its Application in Nonlinear Industrial Process Modeling [J].
Chu, Fei ;
Liang, Tao ;
Chen, C. L. Philip ;
Wang, Xuesong ;
Ma, Xiaoping .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2020, 31 (08) :3017-3031