智能互联基础设施多变量时间序列预测中的对抗性攻击与防御

Pooja Krishan, Rohan Mohapatra, Saptarshi Sengupta
{"title":"智能互联基础设施多变量时间序列预测中的对抗性攻击与防御","authors":"Pooja Krishan, Rohan Mohapatra, Saptarshi Sengupta","doi":"arxiv-2408.14875","DOIUrl":null,"url":null,"abstract":"The emergence of deep learning models has revolutionized various industries\nover the last decade, leading to a surge in connected devices and\ninfrastructures. However, these models can be tricked into making incorrect\npredictions with high confidence, leading to disastrous failures and security\nconcerns. To this end, we explore the impact of adversarial attacks on\nmultivariate time-series forecasting and investigate methods to counter them.\nSpecifically, we employ untargeted white-box attacks, namely the Fast Gradient\nSign Method (FGSM) and the Basic Iterative Method (BIM), to poison the inputs\nto the training process, effectively misleading the model. We also illustrate\nthe subtle modifications to the inputs after the attack, which makes detecting\nthe attack using the naked eye quite difficult. Having demonstrated the\nfeasibility of these attacks, we develop robust models through adversarial\ntraining and model hardening. We are among the first to showcase the\ntransferability of these attacks and defenses by extrapolating our work from\nthe benchmark electricity data to a larger, 10-year real-world data used for\npredicting the time-to-failure of hard disks. Our experimental results confirm\nthat the attacks and defenses achieve the desired security thresholds, leading\nto a 72.41% and 94.81% decrease in RMSE for the electricity and hard disk\ndatasets respectively after implementing the adversarial defenses.","PeriodicalId":501291,"journal":{"name":"arXiv - CS - Performance","volume":"43 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Adversarial Attacks and Defenses in Multivariate Time-Series Forecasting for Smart and Connected Infrastructures\",\"authors\":\"Pooja Krishan, Rohan Mohapatra, Saptarshi Sengupta\",\"doi\":\"arxiv-2408.14875\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The emergence of deep learning models has revolutionized various industries\\nover the last decade, leading to a surge in connected devices and\\ninfrastructures. However, these models can be tricked into making incorrect\\npredictions with high confidence, leading to disastrous failures and security\\nconcerns. To this end, we explore the impact of adversarial attacks on\\nmultivariate time-series forecasting and investigate methods to counter them.\\nSpecifically, we employ untargeted white-box attacks, namely the Fast Gradient\\nSign Method (FGSM) and the Basic Iterative Method (BIM), to poison the inputs\\nto the training process, effectively misleading the model. We also illustrate\\nthe subtle modifications to the inputs after the attack, which makes detecting\\nthe attack using the naked eye quite difficult. Having demonstrated the\\nfeasibility of these attacks, we develop robust models through adversarial\\ntraining and model hardening. We are among the first to showcase the\\ntransferability of these attacks and defenses by extrapolating our work from\\nthe benchmark electricity data to a larger, 10-year real-world data used for\\npredicting the time-to-failure of hard disks. Our experimental results confirm\\nthat the attacks and defenses achieve the desired security thresholds, leading\\nto a 72.41% and 94.81% decrease in RMSE for the electricity and hard disk\\ndatasets respectively after implementing the adversarial defenses.\",\"PeriodicalId\":501291,\"journal\":{\"name\":\"arXiv - CS - Performance\",\"volume\":\"43 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-27\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Performance\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.14875\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Performance","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.14875","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

过去十年间,深度学习模型的出现彻底改变了各行各业,导致互联设备和基础设施激增。然而,这些模型可能会被诱骗,以极高的置信度做出错误的预测,从而导致灾难性的失败和安全问题。为此,我们探讨了对抗性攻击对多变量时间序列预测的影响,并研究了应对方法。具体来说,我们采用了非目标白盒攻击,即快速梯度符号法(FGSM)和基本迭代法(BIM),来毒化训练过程的输入,从而有效地误导模型。我们还展示了攻击后对输入的微妙修改,这使得用肉眼检测攻击变得相当困难。在证明了这些攻击的可行性之后,我们通过对抗训练和模型加固开发出了稳健的模型。通过将我们的工作从基准电力数据外推到用于预测硬盘故障时间的更大的 10 年真实世界数据,我们成为展示这些攻击和防御的可移植性的首批研究者之一。我们的实验结果证实,攻击和防御达到了预期的安全阈值,在实施对抗防御后,电力和硬盘数据集的 RMSE 分别降低了 72.41% 和 94.81%。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
Adversarial Attacks and Defenses in Multivariate Time-Series Forecasting for Smart and Connected Infrastructures
The emergence of deep learning models has revolutionized various industries over the last decade, leading to a surge in connected devices and infrastructures. However, these models can be tricked into making incorrect predictions with high confidence, leading to disastrous failures and security concerns. To this end, we explore the impact of adversarial attacks on multivariate time-series forecasting and investigate methods to counter them. Specifically, we employ untargeted white-box attacks, namely the Fast Gradient Sign Method (FGSM) and the Basic Iterative Method (BIM), to poison the inputs to the training process, effectively misleading the model. We also illustrate the subtle modifications to the inputs after the attack, which makes detecting the attack using the naked eye quite difficult. Having demonstrated the feasibility of these attacks, we develop robust models through adversarial training and model hardening. We are among the first to showcase the transferability of these attacks and defenses by extrapolating our work from the benchmark electricity data to a larger, 10-year real-world data used for predicting the time-to-failure of hard disks. Our experimental results confirm that the attacks and defenses achieve the desired security thresholds, leading to a 72.41% and 94.81% decrease in RMSE for the electricity and hard disk datasets respectively after implementing the adversarial defenses.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
HRA: A Multi-Criteria Framework for Ranking Metaheuristic Optimization Algorithms Temporal Load Imbalance on Ondes3D Seismic Simulator for Different Multicore Architectures Can Graph Reordering Speed Up Graph Neural Network Training? An Experimental Study The Landscape of GPU-Centric Communication A Global Perspective on the Past, Present, and Future of Video Streaming over Starlink
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1