arrow
返回

Sparse Bayesian deep learning for dynamic system identification

delete2022-10-01
delete12
delete
OA
AI
H
Hongpeng Zhou
C
Chahine Ibrahim
Wei Xing Zheng 封面图
Wei Xing Zheng (Wei Xing Zheng)
W
Wei Pan *
DOI:10.1016/j.automatica.2022.110489delete
delete原文链接
delete分享
delete收藏
查看原文
摘要

摘要

En 中文
This paper proposes a sparse Bayesian treatment of deep neural networks (DNNs) for system iden-tification. Although DNNs show impressive approximation ability in various fields, several challenges still exist for system identification problems. First, DNNs are known to be too complex that they can easily overfit the training data. Second, the selection of the input regressors for system identification is nontrivial. Third, uncertainty quantification of the model parameters and predictions are necessary. The proposed Bayesian approach offers a principled way to alleviate the above challenges by marginal likelihood/model evidence approximation and structured group sparsity-inducing priors construction. The identification algorithm is derived as an iterative regularised optimisation procedure that can be solved as efficiently as training typical DNNs. Remarkably, an efficient and recursive Hessian calculation method for each layer of DNNs is developed, turning the intractable training/optimisation process into a tractable one. Furthermore, a practical calculation approach based on the Monte-Carlo integration method is derived to quantify the uncertainty of the parameters and predictions. The effectiveness of the proposed Bayesian approach is demonstrated on several linear and nonlinear system identification benchmarks by achieving good and competitive simulation accuracy. The code to reproduce the experimental results is open-sourced and available online. (C) 2022 Published by Elsevier Ltd.
Keyword:
Regularised system identification
Deep neural networks
Group sparsity
Sparse Bayesian learning
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

Automatica 封面图
Automatica
IF:
5.9
论文数:
1.2W
被引数:
6.5W

机构

D
delft university of technology
学者数:
2.0K
论文数: 825
被引数: 2.3W
W
Western Sydney University
学者数:
661
论文数: 317
被引数: 8.2K
引用论文

引用论文

err分享
err收藏
err分享
err收藏
Variable selection using neural-network models
err2000-03-01
err113
PREAI
errCastellano, G; Fanelli, AM
err分享
err收藏
err分享
err收藏
学者 查看更多内容