计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (1): 89-96.doi: 10.11896/jsjkx.241200190

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于KAN的无监督多元时间序列异常检测网络

王成, 金城   

  1. 复旦大学计算机科学技术学院 上海 200438
  • 收稿日期:2024-12-26 修回日期:2025-03-12 发布日期:2026-01-08
  • 通讯作者: 金城(jc@fudan.edu.cn)
  • 作者简介:(whiletruedo@163.com)
  • 基金资助:
    国家自然科学基金(62472097)

KAN-based Unsupervised Multivariate Time Series Anomaly Detection Network

WANG Cheng, JIN Cheng   

  1. School of Computer Science, Fudan University, Shanghai 200438, China
  • Received:2024-12-26 Revised:2025-03-12 Online:2026-01-08
  • About author:WANG Cheng,born in 1985,Ph.D candidate.His main research interest is time series anomaly detection.
    JIN Cheng,born in 1978,Ph.D,professor,Ph.D supervisor,is a member of CCF(No.08763S).His main research interests include computer vision and multimedia information retrieval.
  • Supported by:
    National Natural Science Foundation of China(62472097).

PDF (PC)

20

摘要： 时间序列数据在金融、医疗、工业和交通等领域中广泛存在,异常检测对确保系统稳定和安全至关重要。由于异常样本的收集十分困难,当前大多数时间序列异常检测方法是无监督的。然而,这些方法普遍存在过泛化问题,即模型不仅能重建正常样本,还能很好地重建异常样本。这一问题使得异常检测效果不佳。因此,提出了一种基于Kolmogorov-Arnold表示理论的时间序列异常检测方法TS-KAN,利用其参数高效性与局部可塑性,使模型更好地拟合正常样本并缓解过泛化问题。此外,提出了局部特征增强层Local-KAN,以增强时域特征的表达能力,提高上下文异常检测能力。在5个主流时间序列异常检测数据集上的实验表明,TS-KAN的异常检测能力显著优于现有方法。

关键词: 时间序列异常检测, KAN, Transformer, 记忆模块, 深度学习

Abstract: Time series data is widely present in fields such as finance,healthcare,industry,and transportation.Time Series Ano-maly Detection(TSAD) is crucial for ensuring system stability and safety.Most current time series anomaly detection methods are unsupervised due to the difficulty in collecting anomaly samples.However,these methods commonly face the problem of over-generalization,where the model can not only reconstruct normal samples,but also effectively reconstruct anomaly samples,leading to poor anomaly detection performance.Therefore,this paper proposes a time series anomaly detection method based on Kolmo-gorov-Arnold representation theory,called TS-KAN.TS-KAN leverages its parameter efficiency and local plasticity to better fit normal samples and alleviate the overgeneralization problem.Additionally,this paper introduces a local feature enhancement layer,namely Local-KAN,to enhance the representation of temporal features and improve contextual anomaly detection capability.Experiments on five mainstream time series anomaly detection datasets demonstrate that TS-KAN significantly outperforms existing methods in anomaly detection capability.

Key words: Time series anomaly detection, KAN, Transformer, Memory module, Deep learning

中图分类号: 

  • TP391
[1]ZONG B,SONG Q,MIN MARTIN R Q,et al.Deep autoenco-ding gaussian mixture model for unsupervised anomaly detection[C]//International Conference on Learning Representations.2018.
[2]YAIRI T,TAKEISHI N,ODA T,et al.A data-driven healthmonitoring method for satellite housekeeping data based on probabilistic clustering and dimensionality reduction[J].IEEE Transactions on Aerospace and Electronic Systems,2017,53(3):1384-1401.
[3]ZHOU Q H,HE S B,LIU H Y,et al.Label-free multivariate time series anomaly detection[J].IEEE Transactions on Know-ledge and Data Engineering,2024,36(7):3166-3179.
[4]FREHNER R B,WU K S,SIM A,et al.Detecting Anomalies in Time Series Using Kernel Density Approaches[J].IEEE Access,2024,12:33420-33439.
[5]SHEN L F,LI Z C,KWOK J.Timeseries anomaly detectionusing temporal hierarchical one-class network[J].Advances in Neural Information Processing Systems,2020,33:13016-13026.
[6]SHIN Y J,LEE S,TARIQ S,et al.Itad:integrative tensor-based anomaly detection system for reducing false positives of satellite systems[C]//Proceedings of the 29th ACM International Conference on Information & Knowledge Management.2020:2733-2740.
[7]DONG C,TAO J F,CHAO Q,et al.Subsequence time series clustering-based unsupervised approach for anomaly detection of axial piston pumps[J].IEEE Transactions on Instrumentation and Measurement,2023,72:1-12.
[8]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.2018:387-395.
[9]TARIQ S,LEE S,SHIN Y J,et al.Detecting anomaliesin space using multivariate convolutional LSTM with mixtures of probabilistic PCA[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.2019:2123-2133.
[10]PARK D,HOSHI Y,KEMP CHARLES C.A multimodal ano-maly detector for robot-assisted feeding using an lstm-based variational autoencoder[J].IEEE Robotics and Automation Letters,2018,3(3):1544-1551.
[11]SU Y,ZHAO Y J,NIU C H,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.2019:2828-2837.
[12]LI Z H,ZHAO Y J,HAN J Q,et al.Multivariate time seriesanomaly detection and interpretation using hierarchical inter-metric and temporal embedding[C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining.2021:3220-3230.
[13]XU J,WU H,WANG J,et al.Anomaly Transformer:Time Series Anomaly Detection with Association Discrepancy[C]//International Conference on Learning Representations.2022.
[14]SUN Y Y,CHEN Z D,FENG C,et al.UMTS-Mixer:Time Series Anomaly Detection Based on Temporal Correlation and Channel Correlation[J].Computer Systems and Applications,2024,33(1):127-133.
[15]YE L,HE Z.Multiscale time series anomaly detection incorporating wavelet decomposition[J].Journal of Computer Applications,2024,44(10):3300-3306.
[16]GONG D,LIU L Q,LE V,et al.Memorizing normality to detect anomaly:Memory-augmented deep autoencoder for unsupervised anomaly detection[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2019:1705-1714.
[17]PARK H J,NOH J,HAM B.Learning memory-guided normality for anomaly detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2020:14372-14381.
[18]SONG J,KIM K,OH J,et al.MEMTO:Memory-guided transformer for multivariate time series anomaly detection[J].arXiv:2312.02530,2023.
[19]LIU Z M,WANG Y X,VAIDYA S,et al.KAN:Kolmogorov-arnold networks[J].arXiv:2404.19756,2024.
[20]LIU Z M,MA P C,WANG Y X,et al.KAN 2.0:Kolmogorov-Arnold Networks Meet Science[J].arXiv:2408.10205,2024.
[21]SIDHARTH S S.Chebyshev polynomial-based kolmogorov-arnold networks:An efficient architecture for nonlinear function approximation[J].arXiv:2405.07200,2024.
[22]AGHAEI A A.fKAN:Fractional Kolmogorov-Arnold Networks with trainable Jacobi basis functions[J].arXiv:2406.07456,2024.
[23]BOZORGASL Z,CHEN H.Wav-KAN:Wavelet kolmogorov-arnold networks[J].arXiv:2405.12832,2024.
[24]LI C X,LIU X Y,LI W Y,et al.U-KAN Makes Strong Backbone for Medical Image Segmentation and Generation[J].arXiv:2406.02918,2024.
[25]VACA-RUBIO C J,BLANCO L,PEREIRA R,et al.Kolmogorov-arnold networks(kans) for time series analysis[J].arXiv:2405.08790,2024.
[26]GENET R,INZIRILLO H.TKAN:Temporal Kolmogorov-Arnold Networks[J].arXiv:2405.07344,2024.
[27]SONODA S,MURATA N.Neural network with unboundedactivation functions is universal approximator[J].Applied and Computational Harmonic Analysis,2017,43(2):233--268.
[28]LAI M J,SHEN Z M.The kolmogorov superposition theoremcan break the curse of dimensionality when approximating high dimensional functions[J].arXiv:2112.09963,2021.
[29]ELFWING S,UCHIBE E,DOYA K.Sigmoid-weighted linearunits for neural network function approximation in reinforcement learning[J].Neural networks,2018,107:3-11.
[30]SUBAKAN C,RAVANELLI M,CORNELL S,et al.Attention is all you need in speech separation[C]//2021 IEEE International Conference on Acoustics,Speech and Signal Processing(ICASSP).2021:21-25.
[31]SARFRAZ M S,CHEN M Y,LAYER L,et al.Position:Quo Vadis,Unsupervised Time Series Anomaly Detection?[C]//Proceedings of the 41st International Conference on Machine Learning.2024:43461-43476.
[32]SU Y,ZHAO Y J,NIU C H,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.2019:2828-2837.
[33]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.2018:387-395.
[34]LI D,CHEN D C,JIN B H,et al.MAD-GAN:Multivariateanomaly detection for time series data with generative adversarial networks[C]//International Conference on Artificial Neural Networks.2019:703-716.
[35]ABDULAAL A,LIU Z H,LANCEWICKI T.Practical approach to asynchronous multivariate time series anomaly detection and localization[C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining.2021:2485-2494.
[36]WU R,KEOGH E J.Current time series anomaly detectionbenchmarks are flawed and are creating the illusion of progress[J].IEEE Transactions on Knowledge and Data Engineering,2021,35(3):2421-2429.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路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