Contrastive-Enhanced Spatial-Temporal Normality Learning for Multivariate Time Series Anomaly Detection and Interpretation

Huangyining Gao; Ruyue Xin; Peng Chen; Xi Li; Ang Bian; Yunni Xia

doi:10.1109/JAS.2026.126194

IEEE/CAA Journal of Automatica Sinica

JCR Impact Factor: 18.3, Top 1 (SCI Q1)

CiteScore: 28.2, Top 1% (Q1)
Google Scholar h5-index: 95， TOP 5

Turn off MathJax

Article Contents

Article Navigation > IEEE/CAA Journal of Automatica Sinica > 2026 > In Press, Accepted Manuscript

H. Gao, R. Xin, P. Chen, X. Li, A. Bian, and Y. Xia, “Contrastive-enhanced spatial-temporal normality learning for multivariate time series anomaly detection and interpretation,” IEEE/CAA J. Autom. Sinica, early access, 2026. doi: 10.1109/JAS.2026.126194

Citation:

H. Gao, R. Xin, P. Chen, X. Li, A. Bian, and Y. Xia, “Contrastive-enhanced spatial-temporal normality learning for multivariate time series anomaly detection and interpretation,” IEEE/CAA J. Autom. Sinica, early access, 2026. doi: 10.1109/JAS.2026.126194

Citation:

H. Gao, R. Xin, P. Chen, X. Li, A. Bian, and Y. Xia, “Contrastive-enhanced spatial-temporal normality learning for multivariate time series anomaly detection and interpretation,” IEEE/CAA J. Autom. Sinica, early access, 2026. doi: 10.1109/JAS.2026.126194

PDF( 2862 KB)

Contrastive-Enhanced Spatial-Temporal Normality Learning for Multivariate Time Series Anomaly Detection and Interpretation

doi: 10.1109/JAS.2026.126194

Funds: This research is supported by the National Key Research and Development Program of China (2024YFF0908200), the National Natural Science Foundation of China (62376043), Sichuan Provincial Natural Science Foundation (2024NSFTD0008), Science and Technology Program of Sichuan Province (2024ZHCG0016), Culture and Tourism Program of of Sichuan Province (2024YB19), Science and Technology Program of Chengdu (2025-GH02-00020-HZ), European Union projects: ENVRI-Hub Next (101131141), EVERSE (101129744), BlueCloud-2026 (101094227), OSCARS (101129751), the China University of Geosciences (Beijing) (2-9-2024-023), the National Science and Technology Major Project (14022025013) and the China University of Geosciences (Beijing) (590125082)

More Information

Author Bio:
Huangyining Gao received his B.E. degree in digital media technology from Yibin University, Yibin, China in 2023. He is currently a master student at the School of Computer and Software Engineering, Xihua University, Chengdu, China. His research interests include machine learning, time series analysis and anomaly detection

Ruyue Xin IEEE member, received her B.E. degree in information management and information system from Communication University of China, Beijing, China in 2012, her M.Sc. degree in system analysis and integration from Beijing Normal University, Beijing, China in 2016 and Ph.D. degree in computer science and technology from the University of Amsterdam, the Netherlands in 2023. She is currently a lecturer at the School of Artificial Intelligence, China University of Geosciences (Beijing). Her research interests include AIOps, anomaly detection, and causal inference

Peng Chen IEEE member, received his B.E. degree in computer science and technology from University of Electronic Science and Technology of China, Chengdu, China in 2001, the M.Sc. degree in computer software and theory from Peking University, Beijing, China in 2004 and Ph.D. degree in computer science and technology from Sichuan University, Chengdu, China in 2017. He is currently a full professor of School of Computer and Software Engineering, Xihua University, Chengdu, China. His research interests include machine learning, anomaly detection and time series analysis

Xi Li received her B.E. (2002), M.Sc. (2005) degree in Computer Science and Technology from Wuhan University of Technology, Wuhan, China and the Ph.D. (2009) degree in the same major from Huazhong University of Science and Technology, Wuhan, China. She is now an Associate Professor at the School of Computer and Software Engineering, Xihua University, Chengdu, China. Her research interests include network and information security, anomaly detection and time series analysis

Ang Bian received her bachelor degree in mathematical statistics in 2010 and the master degree in software engineering in 2013 from Sichuan University, Chengdu, China, and Ph.D. degree in computer science from University of Muenster, Germany in 2018. She is currently a lecturer of School of Computer and Software Engineering, Xihua University, Chengdu, China. Her research interests include computer vision, machine learning and time series analysis

Yunni Xia IEEE Senior member, received his B.S degree in computer science from Chongqing University, in 2003, and his Ph.D. degrees in computer science from Peking University (PKU), in 2008. He is currently a professor in the college of Computer Science at Chongqing University. He is the author or co-author of more than 100 research publications. His research interests are in Petri nets, software quality, performance evaluation, edge computing and cloud computing system dependability
Corresponding author: Peng Chen, e-mail: chenpeng@mail.xhu.edu.cn; Yunni Xia, e-mail: xiayunni@hotmail.com
Received Date: 2026-02-04
Revised Date: 2026-03-27
Accepted Date: 2026-05-06

Available Online: 2026-11-02

Abstract

Abstract

Multivariate time series anomaly detection and interpretation are crucial across domains like finance and system monitoring, where deviations from normal patterns often indicate failures or emerging risks. Although normality learning in a self-supervised manner can effectively capture patterns from an unlabeled time series, current approaches typically focus solely on learning either global or local temporal features and fail to model the normality of a time series from a spatial perspective. Furthermore, a lack of robustness to noise and disturbances can compromise the detection accuracy of these methods in complex scenarios. To address these limitations, we propose CESTaL, a C ontrastive- E nhanced S patio- T emporal norm a lity L earning framework. CESTaL introduces three novel components: 1) A Temporal Order and Relation Learning module (TORL) for temporal pattern consistency through permutation alignment and ranking constraints; 2) A Dynamic Graph Topology Learning (DGTL) module that automatically discovers spatial dependencies via adaptive graph networks; 3) A Contrastive Learning-based Normality Enhancement (CLNE) module learning noise-robust and invariant representations through contrastive learning. Additionally, we identify root cause dimensions through dimension-level reconstruction analysis and cross-view relation divergence. Extensive experiments on six benchmark datasets demonstrate that CESTaL achieves state-of-the-art detection performance, outperforming 10 baseline methods across both point-based and robust range-based evaluation metrics, and also delivers superior results on multiple anomaly explanation metrics, showing strong interpretability.
- Anomaly detection,
- anomaly interpretation,
- contrastive learning,
- data science,
- machine learning,
- multivariate time series,
- normality modeling,
- self-supervise learning

FullText(HTML)

Huangyining Gao and Ruyue Xin contributed equally to this work.

References(46)

References

[1]	M. M. Khan and M. Alkhathami, “Anomaly detection in IoT-based healthcare: Machine learning for enhanced security,” Sci. Rep., vol. 14, no. 1, Art. no. 5872, Mar. 2024. doi: 10.1038/s41598-024-56126-x
[2]	H. Zhu, M. Zhou, Y. Xie, and A. Albeshri, “A self-adapting and efficient dandelion algorithm and its application to feature selection for credit card fraud detection,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 2, pp. 377−390, Feb. 2024. doi: 10.1109/JAS.2023.124008
[3]	T. Chen, C. Zhang, W. Jing, E. Y. S. Foo, L. Sun, and N. Zeng, “Distributed multi-agent fusion state estimation method based on finite-time average consensus for large-scale power systems,” Inf. Fusion, vol. 127, Art. no. 103753, Mar. 2026. doi: 10.1016/j.inffus.2025.103753
[4]	H. Gao, R. Xin, P. Chen, X. Li, N. Lu, and P. You, “Memory-augment graph transformer based unsupervised detection model for identifying performance anomalies in highly-dynamic cloud environments,” J. Cloud Comput., vol. 14, no. 1, Art. no. 40, Jul. 2025. doi: 10.1186/s13677-025-00766-5
[5]	A. Blázquez-García, A. Conde, U. Mori, and J. A. Lozano, “A review on outlier/anomaly detection in time series data,” ACM Comput. Surv., vol. 54, no. 3, Art. no. 56, Apr. 2021.
[6]	J. Fang, Z. Wang, W. Liu, N. Zeng, Y. He, Y. Cao, L. Chen, and X. Liu, “Learning with noisy labels for industrial time series outlier detection: A transformer-embedded contrastive learning framework,” IEEE Trans. Ind. Inform., vol. 22, no. 2, pp. 903−913, Feb. 2026. doi: 10.1109/TII.2025.3616850
[7]	R. Xin, H. Liu, P. Chen, and Z. Zhao, “Robust and accurate performance anomaly detection and prediction for cloud applications: A novel ensemble learning-based framework,” J. Cloud Comput., vol. 12, no. 1, Art. no. 7, Jan. 2023. doi: 10.1186/s13677-022-00383-6
[8]	Y. Chen, H. Xu, G. Pang, H. Qiao, Y. Zhou, and M. Shang, “Self-supervised spatial-temporal normality learning for time series anomaly detection,” in Proc. European Conf. Machine Learning and Knowledge Discovery in Databases. Research Track, Vilnius, Lithuania, 2024, pp. 145−162.
[9]	J. Yang and Z. Yue, “Learning hierarchical spatial-temporal graph representations for robust multivariate industrial anomaly detection,” IEEE Trans. Ind. Inform., vol. 19, no. 6, pp. 7624−7635, Jun. 2023. doi: 10.1109/TII.2022.3216006
[10]	J. Fang, Z. Wang, W. Liu, and X. Liu, “A transfer-learning-assisted role-differentiated approach for industrial outlier detection under label noise,” in Proc. 30th Int. Conf. Automation and Computing, Loughborough, United Kingdom, 2025, pp. 1−6.
[11]	Y. Wang, Q. Wei, H. Li, Y. Du, D. Ni, S. Jiao, W. Zhang, and G. Han, “Toward more generalizable time-series outlier detection in large-volume manufacturing via contextual convolution,” Neurocomputing, vol. 667, Art. no. 132333, Feb. 2026. doi: 10.1016/j.neucom.2025.132333
[12]	S. Naseer, Y. Saleem, S. Khalid, M. K. Bashir, J. Han, M. M. Iqbal, and K. Han, “Enhanced network anomaly detection based on deep neural networks,” IEEE Access, vol. 6, pp. 48231−48246, Aug. 2018. doi: 10.1109/ACCESS.2018.2863036
[13]	A. Sánchez-Ferrera, B. Calvo, and J. A. Lozano, “A review on self-supervised learning in time series anomaly detection: Recent advances and open challenges,” ACM Comput. Surv., vol. 58, no. 5, Art. no. 119, Nov. 2025.
[14]	A. Goodge, B. Hooi, S. K. Ng, and W. S. Ng, “Robustness of autoencoders for anomaly detection under adversarial impact,” in Proc. 30th Int. Joint Conf. Artificial Intelligence, Yokohama, Japan, 2021, Art. no. 173.
[15]	S. Tuli, G. Casale, and N. R. Jennings, “TranAD: Deep transformer networks for anomaly detection in multivariate time series data,” Proc. VLDB Endow., vol. 15, no. 6, pp. 1201−1214, Feb. 2022. doi: 10.14778/3514061.3514067
[16]	S. Schmidl, P. Wenig, and T. Papenbrock, “Anomaly detection in time series: A comprehensive evaluation,” Proc. VLDB Endow., vol. 15, no. 9, pp. 1779−1797, May 2022. doi: 10.14778/3538598.3538602
[17]	R. Xin, P. Chen, and Z. Zhao, “CausalRCA: Causal inference based precise fine-grained root cause localization for microservice applications,” J. Syst. Softw., vol. 203, Art. no. 111724, Sep. 2023. doi: 10.1016/j.jss.2023.111724
[18]	C. Zhang, D. Song, Y. Chen, X. Feng, C. Lumezanu, W. Cheng, J. Ni, B. Zong, H. Chen, and N. V. Chawla, “A deep neural network for unsupervised anomaly detection and diagnosis in multivariate time series data,” in Proc. 33rd AAAI Conf. Artificial Intelligence, Honolulu, USA, 2019, pp. 1409−1416.
[19]	J. Pan, W. Ji, B. Zhong, P. Wang, X. Wang, and J. Chen, “DUMA: Dual mask for multivariate time series anomaly detection,” IEEE Sens. J., vol. 23, no. 3, pp. 2433−2442, Feb. 2023. doi: 10.1109/JSEN.2022.3225338
[20]	G. Li, Z. Yang, H. Wan, and M. Li, “Anomaly-PTG: A time series data-anomaly-detection transformer framework in multiple scenarios,” Electronics, vol. 11, no. 23, Art. no. 3955, Nov. 2022. doi: 10.3390/electronics11233955
[21]	A. Deng and B. Hooi, “Graph neural network-based anomaly detection in multivariate time series,” in Proc. 35th AAAI Conf. Artificial Intelligence, 2021, pp. 4027−4035. (查阅网上资料, 未找到本条文献出版地信息, 请确认)
[22]	P. Qi, D. Li, and S. K. Ng, “MAD-SGCN: Multivariate anomaly detection with self-learning graph convolutional networks,” in Proc. 38th Int. Conf. Data Engineering, Kuala Lumpur, Malaysia, 2022, pp. 1232−1244.
[23]	S. Kim, S. Yun, and J. Kang, “DyGRAIN: An incremental learning framework for dynamic graphs,” in Proc. 31st Int. Joint Conf. Artificial Intelligence, Vienna, Austria, 2022, pp. 3157−3163.
[24]	Q. Zhou, J. Chen, H. Liu, S. He, and W. Meng, “Detecting multivariate time series anomalies with zero known label,” in Proc. 37th AAAI Conf. Artificial Intelligence, Washington, USA, 2023, pp. 4963−4971.
[25]	T. Pranavan, T. Sim, A. Ambikapathi, and S. Ramasamy, “Contrastive predictive coding for anomaly detection in multi-variate time series data,” arXiv preprint arXiv: 2202.03639, 2022.
[26]	B. Zong, Q. Song, M. R. Min, W. Cheng, C. Lumezanu, D. K. Cho, and H. Chen, “Deep autoencoding Gaussian mixture model for unsupervised anomaly detection,” in Proc. 6th Int. Conf. Learning Representations, Vancouver, Canada, 2018.
[27]	S. Lin, R. Clark, R. Birke, S. Schönborn, N. Trigoni, and S. Roberts, “Anomaly detection for time series using VAE-LSTM hybrid model,” in Proc. IEEE Int. Conf. Acoustics, Speech and Signal Processing, Barcelona, Spain, 2020, pp. 4322−4326.
[28]	H. Zhao, Y. Wang, J. Duan, C. Huang, D. Cao, Y. Tong, B. Xu, J. Bai, J. Tong, and Q. Zhang, “Multivariate time-series anomaly detection via graph attention network,” in Proc. IEEE Int. Conf. Data Mining, Sorrento, Italy, 2020, pp. 841−850.
[29]	S. Harford, F. Karim, and H. Darabi, “Generating adversarial samples on multivariate time series using variational autoencoders,” IEEE/CAA J. Autom. Sinica, vol. 8, no. 9, pp. 1523−1538, Sep. 2021. doi: 10.1109/JAS.2021.1004108
[30]	J. Xu, H. Wu, J. Wang, and M. Long, “Anomaly transformer: Time series anomaly detection with association discrepancy,” in Proc. 10th Int. Conf. Learning Representations, 2022. (查阅网上资料, 未找到本条文献出版地信息, 请确认)
[31]	J. Song, K. Kim, J. Oh, and S. Cho, “MEMTO: Memory-guided transformer for multivariate time series anomaly detection,” Proc. 37th Int. Conf. Neural Information Processing Systems, New Orleans, USA, 2023, Art. no. 2525.
[32]	Q. Liu, C. Liu, S. Sun, D. Yao, and Y. Liang, “Gdformer: Going beyond subsequence isolation for multivariate time series anomaly detection,” arXiv preprint arXiv: 2501.18196, 2025.
[33]	W. Zhang, C. Zhang, and F. Tsung, “GRELEN: Multivariate time series anomaly detection from the perspective of graph relational learning,” in Proc. 36th Int. Joint Conf. Artificial Intelligence, Vienna, Austria, 2022, pp. 2390−2397.
[34]	Y. Jeong, E. Yang, J. H. Ryu, I. Park, and M. Kang, “AnomalyBERT: Self-supervised transformer for time series anomaly detection using data degradation scheme,” arXiv preprint arXiv: 2305.04468, 2023.
[35]	J. Liu and S. Chen, “TimesURL: Self-supervised contrastive learning for universal time series representation learning,” in Proc. 38th AAAI Conf. Artificial Intelligence, Vancouver, Canada, 2024, pp. 13918−13926.
[36]	E. Eldele, M. Ragab, Z. Chen, M. Wu, C. K. Kwoh, X. Li, and C. Guan, “Time-series representation learning via temporal and contextual contrasting,” in Proc. 30th Int. Joint Conf. Artificial Intelligence, Montreal, Canada, 2021, pp. 2352−2359.
[37]	Z. Yue, Y. Wang, J. Duan, T. Yang, C. Huang, Y. Tong, and B. Xu, “TS2Vec: Towards universal representation of time series,” in Proc. 36th AAAI Conf. Artificial Intelligence, 2022, pp. 8980−8987. (查阅网上资料, 未找到本条文献出版地信息, 请确认)
[38]	Y. Yang, C. Zhang, T. Zhou, Q. Wen, and L. Sun, “DCdetector: Dual attention contrastive representation learning for time series anomaly detection,” in Proc. 29th ACM SIGKDD Conf. Knowledge Discovery and Data Mining, Long Beach, USA, 2023, pp. 3033−3045.
[39]	Z. Zhong, Z. Yu, X. Xi, Y. Xu, W. Cao, Y. Yang, K. Yang, and J. You, “SimAD: A simple dissimilarity-based approach for time-series anomaly detection,” IEEE Trans. Neural Netw. Learn. Syst., vol. 36, no. 11, pp. 19669−19680, Nov. 2025. doi: 10.1109/TNNLS.2025.3590220
[40]	Z. Z. Darban, G. I. Webb, S. Pan, C. C. Aggarwal, and M. Salehi, “CARLA: Self-supervised contrastive representation learning for time series anomaly detection,” Pattern Recognit., vol. 157, Art. no. 110874, Jan. 2025. doi: 10.1016/j.patcog.2024.110874
[41]	J. Wang, Y. Chen, R. Chakraborty, and S. X. Yu, “Orthogonal convolutional neural networks,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition, Seattle, USA, 2020, pp. 11502−11512.
[42]	S. Li, K. Jia, Y. Wen, T. Liu, and D. Tao, “Orthogonal deep neural networks,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 43, no. 4, pp. 1352−1368, Apr. 2021. doi: 10.1109/TPAMI.2019.2948352
[43]	K. Guo, K. Zhou, X. Hu, Y. Li, Y. Chang, and X. Wang, “Orthogonal graph neural networks,” in Proc. 36th AAAI Conf. Artificial Intelligence, 2022, pp. 3996−4004. (查阅网上资料, 未找到本条文献出版地信息, 请确认)
[44]	A. Zakaria Sellam, I. Benaissa, A. Taleb-Ahmed, L. Patrono, and C. Distante, “MAAT: Mamba adaptive anomaly transformer with association discrepancy for time series,” arXiv preprint arXiv: 2502.07858, 2025.
[45]	A. Huet, J. M. Navarro, and D. Rossi, “Local evaluation of time series anomaly detection algorithms,” in Proc. 28th ACM SIGKDD Conf. Knowledge Discovery and Data Mining, Washington, USA, 2022, pp. 635−645.
[46]	J. Paparrizos, P. Boniol, T. Palpanas, R. S. Tsay, A. Elmore, and M. J. Franklin, “Volume under the surface: A new accuracy evaluation measure for time-series anomaly detection,” Proc. VLDB Endow., vol. 15, no. 11, pp. 2774−2787, Jul. 2022. doi: 10.14778/3551793.3551830

Supplements(0)

Cited By

Proportional views

Proportional views

通讯作者: 陈斌, bchen63@163.com

1.
沈阳化工大学材料科学与工程学院沈阳 110142

Figures(8) / Tables(8)

Get Citation

PDF

XML

Article Metrics

Article views (27) PDF downloads(12)

Contrastive-Enhanced Spatial-Temporal Normality Learning for Multivariate Time Series Anomaly Detection and Interpretation

doi: 10.1109/JAS.2026.126194

Abstract

References

Proportional views

Catalog

通讯作者: 陈斌, bchen63@163.com

Article Metrics

Export File

Citation

Format

Content