Backdoor attack method for load forecasting based on spectral sensitivity identification and sparse graph trigger generator

Zhang Tengfei; Wang Wei; Gao Yuchao; Yang Yang

doi:10.11959/j.issn.2096-109x.AQ25284

您当前的位置：

首页 >

文章列表页 >

Backdoor attack method for load forecasting based on spectral sensitivity identification and sparse graph trigger generator

Papers | 更新时间：2026-07-15

- Backdoor attack method for load forecasting based on spectral sensitivity identification and sparse graph trigger generator
- Chinese Journal of Network and Information Security Vol. 12, Issue 3, Pages: 110-121(2026)
- 作者机构：
  
  南京邮电大学自动化学院，江苏南京 210023
- 作者简介：
- 基金信息：
  
  The National Natural Science Foundation of China(62473204);Key Research Development Plan of Jiangsu Province(BE2023838);Huali Program for Excellent Talents in Nanjing University of Posts and Telecommunications, Postgraduate Research Practice Innovation Program of Jiangsu Province
- DOI：10.11959/j.issn.2096-109x.AQ25284
  CLC： TP309
- Received：24 December 2025，
  
  Revised：2026-04-27，
  
  Accepted：28 April 2026，
  
  Published：25 June 2026
- 稿件说明：
移动端阅览
张腾飞, 王玮, 高煜超, 等. 基于谱敏感性识别与稀疏图触发生成器的负荷预测后门攻击方法[J]. 网络与信息安全学报, 2026, 12(3): 110-121.

Zhang Tengfei, Wang Wei, Gao Yuchao, et al. Backdoor attack method for load forecasting based on spectral sensitivity identification and sparse graph trigger generator[J]. Chinese Journal of Network and Information Security, 2026, 12(3): 110-121.
张腾飞, 王玮, 高煜超, 等. 基于谱敏感性识别与稀疏图触发生成器的负荷预测后门攻击方法[J]. 网络与信息安全学报, 2026, 12(3): 110-121. DOI： 10.11959/j.issn.2096-109x.AQ25284.

Zhang Tengfei, Wang Wei, Gao Yuchao, et al. Backdoor attack method for load forecasting based on spectral sensitivity identification and sparse graph trigger generator[J]. Chinese Journal of Network and Information Security, 2026, 12(3): 110-121. DOI： 10.11959/j.issn.2096-109x.AQ25284.

摘要

在新型电力系统中，基于深度学习的负荷预测模型已被广泛应用于资源调度与市场响应，后门攻击已成为更具隐蔽性的潜在威胁。然而，在这类强耦合、多变量预测场景中，后门触发不仅要实现对预测输出的定向操控，还须兼顾关键样本、变量选择与跨变量依赖。针对上述问题，提出一种基于Koopman谱敏感性识别与图结构扰动建模的后门注入方法KoopTrigger。该方法首先基于Koopman算子敏感性分析选取高脆弱输入窗口以生成高效投毒样本；其次构建门控稀疏图卷积生成器，在变量维度中注入低秩扰动，实现跨窗口的结构化后门触发；最后通过双阶段对抗训练机制稳定诱导模型学习攻击目标。实验在ETTh1、Ireland、Panama与Milan这4个多变量数据集上开展，结果表明KoopTrigger在多个预测模型中均可实现高效、隐蔽的预测操控，性能显著优于消融版本，相较基线方法具有更优的攻击效果与泛化表现。

Abstract

In new-type power systems

deep learning-based load forecasting models have been widely applied to resource scheduling and market response

making backdoor attack a more insidious potential threat. However

in these strongly coupled

multivariate forecasting scenarios

backdoor triggering not only needs to achieve targeted manipulation of the forecast output but also needs to consider key samples

variable selection

and cross-variable dependencies. To address these issues

a backdoor injection method

KoopTrigger

based on Koopman spectral sensitivity identification and graph structure perturbation modeling was proposed. Firstly

this method selects highly vulnerable input windows based on Koopman operator sensitivity analysis to generate efficient poisoning samples; then

it constructs a gated sparse graph convolution generator to inject low-rank perturbations into the variable dimensions

achieving structured backdoor triggering across windows; finally

a two-stage adversarial training mechanism is used to stably induce the model to learn the attack target. Experiments were conducted on four multivariate datasets: ETTh1

Ireland

Panama

and Milan. The results show that KoopTrigger can achieve efficient and covert prediction manipulation in multiple prediction models

significantly outperforming the ablation version. It also shows better attack performance and generalization performance compared to the baseline method.

关键词

Keywords

references

韩富佳 , 王晓辉 , 乔骥 , 等 . 基于人工智能技术的新型电力系统负荷预测研究综述 [J ] . 中国电机工程学报 , 2023 , 43 ( 22 ): 8569 - 8591 .

Han F J , Wang X H , Qiao J , et al . Review on artificial intelligence based load forecasting research for the new-type power system [J ] . Proceedings of the CSEE , 2023 , 43 ( 22 ): 8569 - 8591 .

左宇 , 秦大林 , 杜尔顺 , 等 . 交互式负荷预测: 研究框架与展望 [J ] . 中国电机工程学报 , 2026 , 46 ( 6 ): 2381 - 2394 .

Zuo Y , Qin D L , Du E S , et al . Interactive load forecasting: research framework and outlook [J ] . Proceedings of the CSEE 2026 , 46 ( 6 ): 2381 - 2394 .

Hasan M , Mifta Z , Papiya S J , et al . A state-of-the-art comparative review of load forecasting methods: characteristics, perspectives, and applications [J ] . Energy Conversion and Management: X , 2025 , 26 : 100922 .

Liu M X , Teng F , Zhang Z Y , et al . Enhancing cyber-resiliency of DER-based smart grid: a survey [J ] . IEEE Transactions on Smart Grid , 2024 , 15 ( 5 ): 4998 - 5030 .

《新型电力系统发展蓝皮书》编写组 . 新型电力系统发展蓝皮书 [M ] . 北京：中国电力出版社， 2023 .

The compilation team for the blue book on the development of new-type power systems . Blue book on the development of new-type power systems [M ] . Beijing : China Electric Power Press ， 2023 .

Verma P , Chakraborty C . Load redistribution attacks against smart grids–models, impacts, and defense: a review [J ] . IEEE Transactions on Industrial Informatics , 2024 , 20 ( 8 ): 10192 - 10208 .

Deng W F , Xiang Z L , Huang K K , et al . Detecting intelligent load redistribution attack based on power load pattern learning in cyber-physical power systems [J ] . IEEE Transactions on Industrial Electronics , 2024 , 71 ( 6 ): 6285 - 6293 .

Verma P , Dasgupta P , Chakraborty C . DeepLR: deep learning assisted load redistribution attack on cyber-physical power system [J ] . IEEE Transactions on Smart Grid , 2026 , 17 ( 1 ): 551 - 561 .

Dehbozorgi M R , Rastegar M , Arani M F M . False data injection attack detection and localization framework in power distribution systems using a novel ensemble of CNNs and explainable artificial intelligence [J ] . IEEE Transactions on Industry Applications , 2025 , 61 ( 3 ): 4801 - 4811 .

刘圣龙 , 黄秀丽 , 江伊雯 , 等 . 面向多方数据融合分析的隐私计算技术综述 [J ] . 网络与信息安全学报 , 2024 , 10 ( 6 ): 24 - 36 .

Liu S L , Huang X L , Jiang Y W , et al . Review of privacy computing techniques for multi-party data fusion analysis [J ] . Chinese Journal of Network and Information Security , 2024 , 10 ( 6 ): 24 - 36 .

Ibrahim M S , Dong W , Yang Q . Machine learning driven smart electric power systems: current trends and new perspectives [J ] . Applied Energy , 2020 , 272 : 115237 .

Eren Y , Küçükdemiral İ . A comprehensive review on deep learning approaches for short-term load forecasting [J ] . Renewable and Sustainable Energy Reviews , 2024 , 189 : 114031 .

迟欢欢 , 熊平 , 刘恒竹 , 等 . 激励驱动的联邦学习研究综述: 隐私与安全 [J ] . 网络与信息安全学报 , 2025 , 11 ( 3 ): 19 - 44 .

Chi H H , Xiong P , Liu H Z , et al . Survey on incentive-driven federated learning: privacy and security [J ] . Chinese Journal of Network and Information Security , 2025 , 11 ( 3 ): 19 - 44 .

Li Q , Xu P , He D Q , et al . Multi-source information fusion graph convolution network for traffic flow prediction [J ] . Expert Systems with Applications , 2024 , 252 : 124288 .

Kim N , Park H , Lee J , et al . Short-term electrical load forecasting with multidimensional feature extraction [J ] . IEEE Transactions on Smart Grid , 2022 , 13 ( 4 ): 2999 - 3013 .

Rahman M A , Islam M R , Hossain M A , et al . Resiliency of forecasting methods in different application areas of smart grids: a review and future prospects [J ] . Engineering Applications of Artificial Intelligence , 2024 , 135 : 108785 .

陈晋音 , 瞿康赟 , 郑海斌 . 基于模仿学习的深度强化学习训练数据推断攻击 [J ] . 网络与信息安全学报 , 2025 , 11 ( 5 ): 50 - 67 .

Chen J Y , Qu K Y , Zheng H B . Membership inference attack based on imitation learning for deep reinforcement learning [J ] . Chinese Journal of Network and Information Security , 2025 , 11 ( 5 ): 50 - 67 .

Jiang W Z , Liu H , Liu F . Practical adversarial attacks on spatiotemporal traffic forecasting models [C ] // Proceedings of the 36th International Conference on Neural Information Processing Systems . New Orleans : Curran Associates, Inc. , 2022 : 19035 - 19047 .

Takiddin A , Ismail M , Atat R , et al . Robust graph autoencoder-based detection of false data injection attacks against data poisoning in smart grids [J ] . IEEE Transactions on Artificial Intelligence , 2024 , 5 ( 3 ): 1287 - 1301 .

Fu D Q , Lin X , Liu Z N , et al . BackTime: backdoor attacks on multivariate time series forecasting [C ] // Proceedings of the 38th International Conference on Neural Information Processing Systems . Vancouver : Curran Associates Inc. , 2024 : 131344 - 131268 .

汪旭童 , 尹捷 , 刘潮歌 , 等 . 神经网络后门攻击与防御综述 [J ] . 计算机学报 , 2024 , 47 ( 8 ): 1713 - 1743 .

Wang X T , Yin J , Liu C G , et al . A survey of backdoor attacks and defenses on neural networks [J ] . Chinese Journal of Computers , 2024 , 47 ( 8 ): 1713 - 1743 .

Luo J , Hong T , Fang S C . Benchmarking robustness of load forecasting models under data integrity attacks [J ] . International Journal of Forecasting , 2018 , 34 ( 1 ): 89 - 104 .

Zhang Y , Liu L F , Yuan F M , et al . Multifactor and multiscale method for power load forecasting [J ] . Knowledge-Based Systems , 2023 , 268 : 110476 .

Zhang Q F , Yan H , Liu Y M . Power generation forecasting for solar plants based on dynamic bayesian networks by fusing multi-source information [J ] . Renewable and Sustainable Energy Reviews , 2024 , 202 : 114691 .

Xiang K L , Yang H M , Hao M , et al . BadTime: an effective backdoor attack on multivariate long-term time series forecasting [PP ] . arXiv preprint , 2025 , arXiv: 2508.04189 .

Liu Z X , Liu X L , Xu S H , et al . Beyond immediate activation: temporally decoupled backdoor attacks on time series forecasting [PP ] . arXiv preprint , 2026 , arXiv. 2601 . 04247 .

Chen J Y , Xiong H Y , Zheng H B , et al . Dyn-backdoor: backdoor attack on dynamic link prediction [J ] . IEEE Transactions on Network Science and Engineering , 2024 , 11 ( 1 ): 525 - 542 .

Cai W L , Liang Y X , Liu X G , et al . MSGNet: learning multi-scale inter-series correlations for multivariate time series forecasting [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2024 , 38 ( 10 ): 11141 - 11149 .

Otto S E , Rowley C W . Koopman operators for estimation and control of dynamical systems [J ] . Annual Review of Control, Robotics, and Autonomous Systems , 2021 , 4 : 59 - 87 .

Baddoo P J , Herrmann B , McKeon B J , et al . Physics-informed dynamic mode decomposition [J ] . Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences , 2023 , 479 ( 2271 ): 20220576 .

Jiao R H , Han Z T , Liu X , et al . A gradient-based wind power forecasting attack method considering point and direction selection [J ] . IEEE Transactions on Smart Grid , 2024 , 15 ( 3 ): 3178 - 3192 .

Hu Y Q , Cheng X Y , Wang S H , et al . Times series forecasting for urban building energy consumption based on graph convolutional network [J ] . Applied Energy , 2022 , 307 : 118231 .

Hong J C , Li K R , Liang F W , et al . A novel state of health prediction method for battery system in real-world vehicles based on gated recurrent unit neural networks [J ] . Energy , 2024 , 289 : 129918 .

Zhou H Y . Electricity transformer dataset (ETDataset) [DS ] . GitHub [ 2025-10-23 ] .

Irish social science data archive . ISSDA: Irish social science data archive [DB ] . University College Dublin [ 2025-11-23 ] .

Aguilar Madrid E . Short-term electricity load forecasting (Panama) [DS ] . Kaggle [ 2025-11-01 ] .

Abdullah M . Milan dataset [DS ] . IEEE DataPort .[ 2025-11-01 ] .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Gate-level hardware trojan detection method based on graph Transformer

Research progress on detection of software vulnerabilities based on code similarity

method for machine learning backdoor attacks based on defense method based on diffusion models

Data augmentation based multi-view contrastive learning graph anomaly detection

Multi-query based key node mining algorithm for social networks

Related Author

Zheng Rui

Yu Hong

Xu Yuhang

DONG Weiyu

TANG Yonghe

LIU Chunling

LI Qinghao

ZHANG Aiqing

Related Institution

Information Engineering University

School of Physics and Electronic Information, Anhui Normal University

School of Medical Information, Wannan Medical College

College of Computer and Cyber Security, Fujian Normal University

Fujian Provincial Key Laboratory of Network Security and Cryptology

⁰