计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (7): 237-245.doi: 10.11896/jsjkx.220700078

• 人工智能 • 上一篇    下一篇

上下文信息融合与噪声自适应的异常检测方法

衡红军, 周文华   

  1. 中国民航大学计算机科学与技术学院 天津 300300
  • 收稿日期:2022-07-08 修回日期:2022-12-08 出版日期:2023-07-15 发布日期:2023-07-05
  • 通讯作者: 周文华(zwhhouwhua@163.com)
  • 作者简介:(henghjcauc@163.com)
  • 基金资助:
    国家自然科学基金(U1333109)

Anomaly Detection Method Based on Context Information Fusion and Noise Adaptation

HENG Hongjun, ZHOU Wenhua   

  1. College of Computer Science and Technology,Civil Aviation University of China,Tianjin 300300,China
  • Received:2022-07-08 Revised:2022-12-08 Online:2023-07-15 Published:2023-07-05
  • About author:HENG Hongjun,born in 1968,Ph.D,associate professor,is a member of China Computer Federation.His main research interests include intelligent information processing,natural language,knowledge graph and anomaly detection.ZHOU Wenhua,born in 1998,postgra-duate.Her main research interests include anomaly detection and multiva-riate time series data.
  • Supported by:
    National Natural Science Foundation of China(U1333109).

PDF (PC)

283

摘要： 信息物理系统(CPSs)中传感器和执行器等现场设备收集的数据中隐含复杂的上下文信息和未知分布噪声。为了提取并融合数据中的上下文信息以及减轻噪声带来的干扰,提出了上下文信息融合与噪声自适应的异常检测方法。该方法中设计了一种由自适应降噪编码器、上下文信息编码器和解码器构成的编解码网络建模CPSs状态空间模型。自适应降噪编码器在训练过程中通过拟合数据中噪声的分布模式生成自适应噪声,并利用该噪声对训练数据中的传感器数据加噪,以提升编解码网络的鲁棒性,减轻噪声带来的干扰,同时可迫使降噪自编码器学习到泛化性更强的系统的隐藏状态;上下文信息编码器利用LSTM和CNN提取数据窗口内的时序和空间上下文信息,并使用自注意力机制融合这两类上下文信息和系统隐藏状态,融合结果用于推断当前时刻系统隐藏状态,以提升此隐藏状态中的信息量;解码器利用以上系统隐藏状态可以更准确地解码出相应的传感器数据。编解码网络训练完成后,得到系统隐藏状态和传感器解码值,基于无迹卡尔曼滤波算法计算异常评分。在SWaT和PUMP两个实际CPSs数据集上的实验结果表明,所提方法的F1值均优于其他对比方法,验证了其有效性。

关键词: 异常检测, 自适应噪声, 上下文信息, 状态空间模型, 信息物理系统

Abstract: The data collected by field devices such as sensors and actuators in cyber-physical systems(CPSs) contains complex context information.To extract and fuse context information in data and reduce the interference caused by noise,an anomaly detection method based on context information fusion and noise adaptation is proposed.In this method,an encoder-decoder network composed of adaptive denoising encoder,context information encoder and decoder is designed to model the state-space model of CPSs.The adaptive denoising encoder generates adaptive noise by fitting the distribution of noise in the data during the training process,and adds the adaptive noise to the sensor data of the training data,so as to improve the robustness of the encoder-decoder network,reduce the interference caused by noise,and force the adaptive denoising decoder to learn the system hidden state with stronger generalization.Context information encoder uses long-short term memory(LSTM) and convolutional neural networks(CNN) to extract temporal context information and spatial context information in the data window,and uses self-attention mechanism to fuse these two types of context information with system hidden state,so as to obtain the fusion result,which is used to infer the system hidden state at the current moment,so as to increase the amount of information of system hidden state at the current moment.The decoder can decode the corresponding sensor data more accurately by using the above system hidden states.After the encoder-decoder network training is completed,the system hidden state and the decoded sensor data are obtained,and the anomaly score is calculated based on the unscented Kalman filter algorithm.Experimental results on two actual CPSs datasets of SWaT and PUMP show that the F1 value of the proposed method is better than other comparison methods,which verifies its effectiveness.

Key words: Anomaly detection, Adaptive noise, Context information, State space model, Cyber-physical systems

中图分类号: 

  • TP183
[1]YANG P,STANKEVICIUS D,MAROZAS V,et al.Lifelogging data validation model for internet of things enabled personalized healthcare[J].IEEE Transactions on Systems,Man,and Cybernetics,Systems,2016,48(1):50-64.
[2]YIN C,XI J,SUN R,et al.Location privacy protection based on differential privacy strategy for big data in industrial internet of things[J].IEEE Transactions on Industrial Informatics,2017,14(8):3628-3636.
[3]BLÁZQUEZ-GARCÍA A,CONDE A,MORI U,et al.A review on outlier/anomaly detection in time series data[J].ACM Computing Surveys(CSUR),2021,54(3):1-33.
[4]LIN S,CLARK R,BIRKE R,et al.Anomaly detection for time series using vae-lstm hybrid model[C]//Proceedings of the ICASSP 2020-2020 IEEE International Conference on Acoustics,Speech and Signal Processing(ICASSP).USA,IEEE,2020:4322-4326.
[5]LI J,GAO H.Survey on sensor network research[J].Journal of Computer Research and Development,2008,45(1):1-15.
[6]FEI H,XIAO F,LI G H,et al.An anomaly detection method of wireless sensor network based on multi-modals data stream[J].Chinese Journal of Computers,2017,40(8):1829-1842.
[7]ZHANG C,SONG D,CHEN Y,et al.A deep neural network for unsupervised anomaly detection and diagnosis in multivariate time series data[C]//Proceedings of the AAAI Conference on Artificial Intelligence.USA,AAAI,2019,33(1):1409-1416.
[8]KIM T Y,CHO S B.Web traffic anomaly detection using C-LSTM neural networks[J].Expert Systems with Applications,2018,106:66-76.
[9]YIN C,ZHANG S,WANG J,et al.Anomaly detection based on convolutional recurrent autoencoder for IoT time series[J].IEEE Transactions on Systems,Man,and Cybernetics,Systems,2020,52(1):112-122.
[10]MALHOTRA P,RAMAKRISHNAN A,ANAND G,et al.LSTM-based encoder-decoder for multi-sensor anomaly detection[J].arXiv:1607.00148,2016.
[11]PARK D,HOSHI Y,KEMP C C.A multimodal anomaly detector for robot-assisted feeding using an lstm-based variational autoencoder[J].IEEE Robotics and Automation Letters,2018,3(3):1544-1551.
[12]LI D,CHEN D,JIN B,et al.MAD-GAN,Multivariate anomaly detection for time series data with generative adversarial networks[C]//Proceedings of International Conference on Artificial Neural Networks.Cham,Springer,2019:703-716.
[13]SCHREYER M,SATTAROV T,SCHULZE C,et al.Detection of accounting anomalies in the latent space using adversarial autoencoder neural networks[J].arXiv:1908.00734,2019.
[14]AUDIBERT J,MICHIARDI P,GUYARD F,et al.Usad,Unsupervised anomaly detection on multivariate time series[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.USA,ACM,2020:3395-3404.
[15]HUNDMAN K,CONSTANTINOU V,LAPORTE C,et al.Detecting spacecraft anomalies using lstms and nonparametric dynamic thresholding[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mi-ning.USA,ACM,2018:387-395.
[16]SU Y,ZHAO Y,NIU C,et al.Robust anomaly detection for multivariate time series through stochastic recurrent neural network[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.USA,ACM,2019:2828-2837.
[17]DENG A,HOOI B.Graph neural network-based anomaly detection in multivariate time series[C]//Proceedings of the AAAI Conference on Artificial Intelligence.USA,AAAI,2021,35(5):4027-4035.
[18]FENG C,TIAN P.Time series anomaly detection for cyber-physical systems via neural system identification and bayesian filtering[C]//Proceedings of the 27th ACM SIGKDD Confe-rence on Knowledge Discovery & Data Mining.USA:ACM,2021:2858-2867.
[19]SCHNEIDER T,QIU C,KLOFT M,et al.Detecting Anomalies within Time Series using Local Neural Transformations[J].arXiv:2202.03944,2022.
[20]LEE E A,SESHIA S A.Introduction to embedded systems,A cyber-physical systems approach[M].USA,MIT Press,2016:181-207.
[21]JULIER S J,UHLMANN J K.Unscented filtering and nonli-near estimation[J].Proceedings of the IEEE,2004,92(3):401-422.
[22]INOUE H.Multi-sample dropout for accelerated training andbetter generalization[J].arXiv:1905.09788,2019.
[23]JULIER S J.The scaled unscented transformation[C]//Proceedings of the 2002 American Control Conference(IEEE Cat.No.CH37301).USA,IEEE,2002:4555-4559.
[24]SIPPLE J.Interpretable,multidimensional,multimodal anomaly detection with negative sampling for detection of device failure[C]//Proceeding of the 37th International Conference on Machine Learning.USA,ACM,2020:9016-9025.
[25]LIU F T,TING K M,ZHOU Z H,et al.Isolation forest[C]//Proceedings of 2008 Eighth IEEE International Conference on Data Mining.USA,IEEE Computer Society,2008:413-422.
[26]NG A.Sparse autoencoder[J].CS294A Lecture Notes,2011,72(2011):1-19.
[27]GOH J,ADEPU S,TAN M,et al.Anomaly detection in cyber physical systems using recurrent neural networks[C]//Procee-dings of 2017 IEEE 18th International Symposium on High Assurance Systems Engineering(HASE).USA,IEEE,2017:140-145.
[28]ZONG B,SONG Q,MIN M R,et al.Deep autoencoding Gaussian mixture model for unsupervised anomaly detection[C]//Proceedings of International Conference on Learning Representations.USA,IEEE,2018:1-19.
[29]SHEN L,LI Z,KWOK J.Timeseries anomaly detection using temporal hierarchical one-class network[J].Advances in Neural Information Processing Systems,2020,33:13016-13026.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发