Semi-asynchronous personalized federated learning for short-term photovoltaic power forecasting

doi:10.1016/j.dcan.2022.03.022

CORC > 自动化研究所 > 中国科学院自动化研究所 > 多模态人工智能系统全国重点实验室

	Semi-asynchronous personalized federated learning for short-term photovoltaic power forecasting
	Zhang, Weishan 1; Chen, Xiao 1; He, Ke 2; Chen, Leiming 1; Xu, Liang 3; Wang, Xiao 4; Yang, Su 5
刊名	DIGITAL COMMUNICATIONS AND NETWORKS
	2023-10-01
卷号	9 期号:5 页码:1221-1229
关键词	Photovoltaic power forecasting Federated learning Edge computing CNN-LSTM
ISSN号	2468-5925
DOI	10.1016/j.dcan.2022.03.022
通讯作者	Zhang, Weishan(zhangws@upc.edu.cn)
英文摘要	Accurate forecasting for photovoltaic power generation is one of the key enablers for the integration of solar photovoltaic systems into power grids. Existing deep-learning-based methods can perform well if there are sufficient training data and enough computational resources. However, there are challenges in building models through centralized shared data due to data privacy concerns and industry competition. Federated learning is a new distributed machine learning approach which enables training models across edge devices while data reside locally. In this paper, we propose an efficient semi-asynchronous federated learning framework for short-term solar power forecasting and evaluate the framework performance using a CNN-LSTM model. We design a personalization technique and a semi-asynchronous aggregation strategy to improve the efficiency of the proposed federated forecasting approach. Thorough evaluations using a real-world dataset demonstrate that the federated models can achieve significantly higher forecasting performance than fully local models while protecting data privacy, and the proposed semi-asynchronous aggregation and the personalization technique can make the forecasting framework more robust in real-world scenarios.
资助项目	National Natural Science Foundation of China[62072469] ; National Key R & D Program of China[2018AAA0101502] ; Shandong Natural Science Foundation[ZR2019MF049] ; West Coast artificial intelligence technology innovation center[2019-1-5] ; West Coast artificial intelligence technology innovation center[2019-1-6] ; Opening Project of Shanghai Trusted Industrial Control Platform[TICPSH202003015-ZC]
WOS研究方向	Telecommunications
语种	英语
出版者	KEAI PUBLISHING LTD
WOS记录号	WOS:001107446900001
资助机构	National Natural Science Foundation of China ; National Key R & D Program of China ; Shandong Natural Science Foundation ; West Coast artificial intelligence technology innovation center ; Opening Project of Shanghai Trusted Industrial Control Platform
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/55231]
专题	多模态人工智能系统全国重点实验室
通讯作者	Zhang, Weishan
作者单位	1.China Univ Petr East China, Coll Comp Sci & Technol, Dongying, Peoples R China 2.Tsinghua Univ, Sichuan Energy Internet Res Inst, Beijing, Peoples R China 3.Beijing Univ Sci & Technol, Sch Comp & Commun Engn, Beijing, Peoples R China 4.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing, Peoples R China 5.Fudan Univ, Sch Comp Sci, Shanghai, Peoples R China
推荐引用方式 GB/T 7714	Zhang, Weishan,Chen, Xiao,He, Ke,et al. Semi-asynchronous personalized federated learning for short-term photovoltaic power forecasting[J]. DIGITAL COMMUNICATIONS AND NETWORKS,2023,9(5):1221-1229.
APA	Zhang, Weishan.,Chen, Xiao.,He, Ke.,Chen, Leiming.,Xu, Liang.,...&Yang, Su.(2023).Semi-asynchronous personalized federated learning for short-term photovoltaic power forecasting.DIGITAL COMMUNICATIONS AND NETWORKS,9(5),1221-1229.
MLA	Zhang, Weishan,et al."Semi-asynchronous personalized federated learning for short-term photovoltaic power forecasting".DIGITAL COMMUNICATIONS AND NETWORKS 9.5(2023):1221-1229.