基于纵向联邦学习的能源排放跨界智能分析

doi:10.19678/j.issn.1000-3428.0068452

摘要/Abstract

摘要：

在企业生产过程中的能源排放预测一直是企业管理和政府监督重点关注的问题, 随着信息采集能力增强, 在能源排放预测过程中会涉及越来越多的跨界数据, 使得预测模型面临着数据量庞大和数据关联性较低的挑战, 从而增大模型的训练难度, 降低预测的准确性。为此, 提出基于纵向联邦学习的能源排放智能预测模型。针对跨领域联合建模过程中数据源分散、信息密度小的问题, 设计基于纵向联邦学习的异步网络更新方法, 保证本地数据的安全和多方建模的质量。异步网络更新方法还能降低多方建模的时间和空间开销。针对模型间通信数据的安全高效传递问题, 设计基于同态加密的数据跨平台通信算法, 利用数据加密在保障通信网络安全的同时使用数据压缩技术减小加密数据的体积, 进一步提高模型间的通信效率。实验结果表明, 该模型具有良好的性能, 相比于基准模型, 所提的能源排放预测模型的R2值最多提升了16%, 能够降低约40%的联合建模时间, 充分证明能源排放跨界智能分析模型解决了跨界数据难以共享共用的问题, 并且提高了跨界联合建模的速度和准确率。

关键词: 能源排放预测, 跨界融合, 纵向联邦学习, 多方建模, 数据安全

Abstract:

Energy emission prediction in enterprise production processes has long been a focal point for enterprise management and government supervision. However, with the increasing use of cross-border data in the prediction process due to enhanced information collection capabilities, prediction models face challenges such as large data volumes and low data relevance, leading to difficulties in training and reduced prediction accuracy. To address these issues, this study proposes an intelligent prediction model for energy emissions based on Vertical Federated Learning (VFL). To handle scattered data sources and low information density during cross-domain joint modeling, an asynchronous network update method based on VFL is introduced, ensuring the safety of local data and the quality of multi-party modeling. Asynchronous network update method can also reduce the time and space overhead of multi-party modeling Additionally, a data cross-platform communication algorithm based on Homomorphic Encryption (HE) is proposed to achieve secure and efficient communication between models. This algorithm employs data encryption for secure communication and data compression techniques to enhance communication efficiency. Experimental results demonstrate that the proposed model outperforms traditional methods, achieving up to a 16% R2 improvement and approximately 40% reduction in joint modeling time. This cross-border intelligent analysis model effectively addresses the sharing and utilization issues of cross-border data, enhancing the speed and accuracy of cross-border joint modeling.

Key words: energy emission prediction, cross-border fusion, Vertical Federated Learning(VFL), multi-party modeling, data security

王圆圆, 王世谦, 王涵, 郭正宾, 胡显承. 基于纵向联邦学习的能源排放跨界智能分析[J]. 计算机工程, 2025, 51(1): 164-173.

WANG Yuanyuan, WANG Shiqian, WANG Han, GUO Zhengbin, HU Xiancheng. Cross-Border Intelligent Analysis of Energy Emission Based on Vertical Federated Learning[J]. Computer Engineering, 2025, 51(1): 164-173.

https://www.ecice06.com/CN/Y2025/V51/I1/164

图/表 10

图1 基于纵向联邦学习的能源排放智能分析模型

Fig.1 Energy emission intelligent analysis model based on vertical federated learning

图2 DIC算法流程

Fig.2 Procedure of DIC algorithm

图3 在不同被动方数量下的安全性对比

Fig.3 Comparison of safety under different numbers of passive parties

图4 不同网络更新方法的单轮更新耗时对比

Fig.4 Comparison of single round update time for different network update methods

图5 不同网络更新方法的模型收敛速度对比

Fig.5 Comparison of convergence speed of models using different network update methods

图6 在不同参与方数量下的损失值对比

Fig.6 Comparison of loss values under different numbers of participants

图7 不同模型的损失值对比

Fig.7 Comparison of loss values for different models

图8 本文模型的排放量预测结果

Fig.8 Emission forecast results of the proposed model

图9 能源排放预测模型R2对比

Fig.9 R2 comparison of energy emission prediction models

参考文献 25

1	冯宗宪, 王安静. 中国区域碳峰值测度的思考和研究——基于全国和陕西省数据的分析. 西安交通大学学报(社会科学版), 2016, 36 (4): 96- 104. doi: 10.15896/j.xjtuskxb.201604014
	FENG Z X , WANG A J . Comparative study of China regional carbon peak-based on national data and Shaanxi province. Journal of Xi'an Jiaotong University (Social Sciences), 2016, 36 (4): 96- 104. doi: 10.15896/j.xjtuskxb.201604014
2	Intergovernmental Panel on Climate Change . Climate change 2013: the physical science basis. New York, USA: Cambridge University Press, 2013.
3	姜克隽, 贺晨旻, 庄幸, 等. 我国能源活动CO2排放在2020—2022年之间达到峰值情景和可行性研究. 气候变化研究进展, 2016, 12 (3): 167- 171. doi: 10.12006/j.issn.1673-1719.2015.200
	JIANG K J , HE C M , ZHUANG X , et al. Scenario and feasibility study for peaking CO2 emission from energy activities in China. Climate Change Research, 2016, 12 (3): 167- 171. doi: 10.12006/j.issn.1673-1719.2015.200
4	SAIDUR R . Energy consumption, energy savings, and emission analysis in Malaysian office buildings. Energy Policy, 2009, 37 (10): 4104- 4113. doi: 10.1016/j.enpol.2009.04.052
5	KHAN N , YAQOOB I , HASHEM I A T , et al. Big data: survey, technologies, opportunities, and challenges. The Scientific World Journal, 2014, 7, 712826. doi: 10.1155/2014/712826
6	LI T , SAHU A K , TALWALKAR A , et al. Federated learning: challenges, methods, and future directions. IEEE Signal Processing Magazine, 2020, 37 (3): 50- 60. doi: 10.1109/MSP.2020.2975749
7	郑美光, 杨泳. 基于互信息软聚类的个性化联邦学习算法. 计算机工程, 2023, 49 (8): 20- 28.
	ZHENG M G , YANG Y . Personalized federated learning algorithm based on mutual information and soft clustering. Computer Engineering, 2023, 49 (8): 20- 28.
8	杨文琦, 章阳, 聂江天, 等. 基于联邦学习的无线网络节点能量与信息管理策略. 计算机工程, 2022, 48 (1): 188-196, 203. doi: 10.19678/j.issn.1000-3428.0060825
	YANG W Q , ZHANG Y , NIE J T , et al. Energy and information management strategy based on federated learning for wireless network nodes. Computer Engineering, 2022, 48 (1): 188-196, 203. doi: 10.19678/j.issn.1000-3428.0060825
9	JORDAN M I , MITCHELL T M . Machine learning: trends, perspectives, and prospects. Science, 2015, 349 (6245): 255- 260. doi: 10.1126/science.aaa8415
10	JANIESCH C , ZSCHECH P , HEINRICH K . Machine learning and deep learning. Electronic Markets, 2021, 31 (3): 685- 695. doi: 10.1007/s12525-021-00475-2
11	LU L L , ZHOU L , ZHANG H Y , et al. The effects of industrial energy consumption on energy-related carbon emissions at national and provincial levels in China. Energy Science[WT《Times New Roman》]& Engineering, 2018, 6 (5): 371- 384.
12	WANG P , ZHONG Y Y , YAO Z N . Modeling and estimation of CO2 emissions in China based on artificial intelligence. Computational Intelligence and Neuroscience, 2022, 32, 6822467.
13	POUYANFAR S , SADIQ S , YAN Y L , et al. A survey on deep learning: algorithms, techniques, and applications. ACM Computing Surveys, 2019, 51 (5): 1- 36. URL
14	QIAO Y C , LAN Q J , ZHOU Z D , et al. Privacy-preserving credit evaluation system based on blockchain. Expert Systems with Applications, 2022, 188, 115989. doi: 10.1016/j.eswa.2021.115989
15	VAN DIJK M, GENTRY C, HALEVI S, et al. Fully homomorphic encryption over the integers[C]//Proceedings of the 29th Annual International Conference on the Theory and Applications of Cryptographic Techniques. Berlin, Germany: Springer, 2010: 24-43.
16	CHAI D , WANG L Y , CHEN K , et al. Secure federated matrix factorization. IEEE Intelligent Systems, 2021, 36 (5): 11- 20. doi: 10.1109/MIS.2020.3014880
17	HUANG W , LI T R , WANG D X , et al. Fairness and accuracy in horizontal federated learning. Information Sciences, 2022, 589, 170- 185. doi: 10.1016/j.ins.2021.12.102
18	MCMAHAN H B, MOORE E, RAMAGE D, et al. Communication-efficient learning of deep networks from decentralized data[EB/OL]. [2023-08-18]. https://arxiv.org/pdf/1602.05629.
19	MUHAMMAD K, WANG Q Q, O'REILLY-MORGAN D, et al. FedFast: going beyond average for faster training of federated recommender systems[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery[WT《Times New Roman》]& Data Mining. New York, USA: ACM Press, 2020: 1234-1242.
20	LI S, LI W Q, COOK C, et al. Independently recurrent neural network (IndRNN): building a longer and deeper RNN[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 5457-5466.
21	CORTES P , RODRIGUEZ J , SILVA C , et al. Delaycompensation in model predictive current control of a three-phase inverter. IEEE Transactions on Industrial Electronics, 2012, 59 (2): 1323- 1325.
22	ACAR A , AKSU H , ULUAGAC A S , et al. A survey on homomorphic encryption schemes. ACM Computing Surveys, 2019, 51 (4): 1- 35.
23	ZHENG S X, MENG Q, WANG T F, et al. Asynchronous stochastic gradient descent with delaycompensation[EB/OL]. [ 2023-08-18]. https://arxiv.org/pdf/1609.08326.
24	AJI A F, HEAFIELD K. Sparsecommunication for distributed gradient descent[EB/OL]. [2023-08-18]. https://arxiv.org/pdf/1704.05021.
25	CHENG H T, KOC L, HARMSEN J, et al. Wide[WT《Times New Roman》]& deep learning for recommender systems[C]//Proceedings of the 1st Workshop on Deep Learning for Recommender Systems. New York, USA: ACM Press, 2016: 7-10.

[1]	熊世强, 何道敬, 王振东, 杜润萌. 联邦学习及其安全与隐私保护研究综述[J]. 计算机工程, 2024, 50(5): 1-15.
[2]	张利华, 曹宇, 张赣哲, 黄阳, 陈世宏. 基于区块链的微电网数据安全共享方案[J]. 计算机工程, 2022, 48(1): 43-50.
[3]	周远林, 陶洋, 李正阳, 杨柳. 基于双簇头的无线传感器网络反馈信任模型[J]. 计算机工程, 2021, 47(3): 174-182.
[4]	郭威, 谢光伟, 张帆, 李敏. 一种分布式存储系统拟态化架构设计与实现[J]. 计算机工程, 2020, 46(6): 12-19.
[5]	胡子杰,张帆,沈继忠,赵新杰. 针对民用无人机的遥控数据保护方法[J]. 计算机工程, 2019, 45(4): 100-107.
[6]	劳卫伦, 王柏勇, 张锐, 王加贝. 基于区块链的电力数据容灾备份方案[J]. 计算机工程, 2019, 45(11): 9-15.
[7]	张悦,于佳. 基于身份的云存储完整性检测方案[J]. 计算机工程, 2018, 44(3): 8-12,18.
[8]	张立红,陈晶,杜瑞颖,何琨,陈炯. 云存储中支持安全去重的数据完整性验证方法[J]. 计算机工程, 2017, 43(1): 32-36,42.
[9]	刘利方,罗亚辉,康江. 应用于AMR的Flash文件系统设计与实现[J]. 计算机工程, 2016, 42(2): 66-69,76.
[10]	王惠峰,李战怀,张晓,孙鉴,赵晓南. 支持并发更新的云存储数据持有性审计方法[J]. 计算机工程, 2016, 42(2): 15-20.
[11]	王慧琳,颜湘涛. 不含双线性对的密文检索公钥加密方案[J]. 计算机工程, 2014, 40(8): 106-111,115.
[12]	刘婷婷, 赵勇. 一种隐私保护的多副本完整性验证方案[J]. 计算机工程, 2013, 39(7): 55-58.
[13]	李冠朋, 田振川, 朱贵良. 基于ECDH与Rijndael的数据库加密系统[J]. 计算机工程, 2013, 39(4): 173-176,179.
[14]	王小威, 赵一鸣. 一种基于任务角色的云计算访问控制模型[J]. 计算机工程, 2012, 38(24): 9-13.
[15]	左黎明, 汤鹏志, 刘二根, 徐保根. 基于行为特征的恶意代码检测方法[J]. 计算机工程, 2012, 38(2): 129-131.

选择文件类型/文献管理软件名称

选择包含的内容