The DOI matches its Crossref record; 19 of the 20 references it cites were found in a registry or an index.
USAD: UnSupervised Anomaly Detection on Multivariate Time Series
Open accessAbstract
The automatic supervision of IT systems is a current challenge at Orange. Given the size and complexity reached by its IT operations, the number of sensors needed to obtain measurements over time, used to infer normal and abnormal behaviors, has increased dramatically making traditional expert-based supervision methods slow or prone to errors. In this paper, we propose a fast and stable method called UnSupervised Anomaly Detection for multivariate time series (USAD) based on adversely trained autoencoders. Its autoencoder architecture makes it capable of learning in an unsupervised way. The use of adversarial training and its architecture allows it to isolate anomalies while providing fast training. We study the properties of our methods through experiments on five public datasets, thus demonstrating its robustness, training speed and high anomaly detection performance. Through a feasibility study using Orange's proprietary data we have been able to validate Orange's requirements on scalability, stability, robustness, training speed and high performance.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“It is beneficial to introduce an adversarial-based SSL, which can further amplify the difference between normal and anomalous data, such as USAD [66] and DAEMON [184].”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [66]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The core idea behind USAD is to amplify the reconstruction error by adversarial training between two BAEs.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [66]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“USAD [66] is a time series anomaly detection framework that includes two BAE models, and two BAE are defined as AE 1 and AE 2 , respectively.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [66]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Adversarial-based regularization constraints can amplify the score difference between anomalies and normal points, thereby enhancing the accuracy of anomalies identification from another perspective, such as Minmax learning [66], [163] and adversarial learning [184].”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [66]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“RANSynCoders [63], USAD[66], AnomalyTrans [163], and DAEMON [184] employ autoencoder-based reconstruction SSL framework.”
From Self-Supervised Learning for Time Series Analysis: Taxonomy, Progress, and Prospects (2024) · cites this paper as [66]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“poral data, methods based on recurrent neural networks [7] (RNNs) are very popular.”
MethodGAN(Generative Adversarial Nets)
(2017) · Journal of Japan Society for Fuzzy Theory and Intelligent Informatics · doi:10.3156/jsoft.29.5_177_2 · Cited by 19,270
Learning Internal Representations by Error Propagation
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1985) · doi:10.21236/ada164453 · Cited by 16,172
Liu, F. T., Ting, K. M., & Zhou, Z.-H. (2008) · 2008 Eighth IEEE International Conference on Data Mining · doi:10.1109/icdm.2008.17 · Cited by 6,601 · cited 1×
“Figure 2(C) presents the results for m ∈ [ 5 , 10 , 20 , 40 , 100 ] . dimension for Z causes a large loss of information at the encoding .”
Background5th International Conference on Learning Representations (ICLR 17)
Havrylov, S., & Titov, I. (2017) · Cited by 3,503
Robust Anomaly Detection for Multivariate Time Series through Stochastic Recurrent Neural Network
Su, Y., Zhao, Y., Niu, C., Liu, R., Sun, W., & Pei, D. (2019) · Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3292500.3330672 · Cited by 1,804 · cited 9×
“On top, the results obtained with SWaT and WADI datasets are presented, whereas the bottom part of the table reports obtained results from the benchmark proposed by [17], using three remaining datasets.”
BackgroundMethodResultInfluentialDetecting Spacecraft Anomalies Using LSTMs and Nonparametric Dynamic Thresholding
Hundman, K., Constantinou, V., Laporte, C., Colwell, I., & Soderstrom, T. (2018) · Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3219819.3219845 · Cited by 1,771
Schwartz, R., Dodge, J., Smith, N. A., & Etzioni, O. (2020) · Communications of the ACM · doi:10.1145/3381831 · Cited by 1,647
Deep Learning for Anomaly Detection: A Survey
Chalapathy, R., & Chawla, S. (2019) · arXiv · doi:10.48550/arxiv.1901.03407 · Cited by 1,240 · cited 1×
“Detecting unexpected behavior on a set of measurements correlated with each other over time is an active research discipline called anomaly detection in multivariate time series [2].”
BackgroundA Multimodal Anomaly Detector for Robot-Assisted Feeding Using an LSTM-Based Variational Autoencoder
Park, D., Hoshi, Y., & Kemp, C. C. (2018) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2018.2801475 · Cited by 1,115
Research works printed but not linked: 3
- Ian J. Goodfellow , Jean Pouget-Abadie , Mehdi Mirza , Bing Xu , David Warde-Farley , Sherjil Ozair , Aaron C. Courville , and Yoshua Bengio . 2014 . Generative Adversarial Nets. In Advances in Neural Information Processing Systems 27: Annual Conference on Neural Information Processing Systems 2014. Montreal, Quebec, Canada, 2672--2680. Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron C. Courville, and Yoshua Bengio. 2014. Generative Adversarial Nets. In Advances in Neural Information Processing Systems 27: Annual Conference on Neural Information Processing Systems 2014. Montreal, Quebec, Canada, 2672--2680.
- Chuxu Zhang , Dongjin Song , Yuncong Chen , Xinyang Feng , Cristian Lumezanu , Wei Cheng , Jingchao Ni , Bo Zong , Haifeng Chen , and Nitesh V. Chawla . 2019. A Deep Neural Network for Unsupervised Anomaly Detection and Diagnosis in Multivariate Time Series Data . Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 33 ( 2019 ), 1409--1416. Chuxu Zhang, Dongjin Song, Yuncong Chen, Xinyang Feng, Cristian Lumezanu, Wei Cheng, Jingchao Ni, Bo Zong, Haifeng Chen, and Nitesh V. Chawla. 2019. A Deep Neural Network for Unsupervised Anomaly Detection and Diagnosis in Multivariate Time Series Data. Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33 (2019), 1409--1416.
- Bo Zong , Qi Song , Martin Renqiang Min , Wei Cheng , Cristian Lumezanu , Daeki Cho , and Haifeng Chen . 2018 . Deep autoencoding Gaussian mixture model for unsupervised anomaly detection . In 6th International Conference on Learning Representations, ICLR 2018. Toulon, France, 1--19. Bo Zong, Qi Song, Martin Renqiang Min, Wei Cheng, Cristian Lumezanu, Daeki Cho, and Haifeng Chen. 2018. Deep autoencoding Gaussian mixture model for unsupervised anomaly detection. In 6th International Conference on Learning Representations, ICLR 2018. Toulon, France, 1--19.
Cited by
Items 101–200 of 1,000
Huang, J., Zeng, G., Geng, G., Weng, J., & Lu, K. (2023) · IET Cyber-Systems and Robotics · doi:10.1049/csy2.12085 · Cited by 35
Cao, Y., Tang, X., Deng, X., & Wang, P. (2024) · Process Safety and Environmental Protection · doi:10.1016/j.psep.2024.04.012 · Cited by 32
Liu, H., Huang, X., Jia, M., Jia, T., Han, J., Wu, Z., & Li, Y. (2024) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2024.3411481 · Cited by 35
Yao, Y., Ma, J., & Ye, Y. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.107757 · Cited by 33
Kong, L., Yu, J., Tang, D., Song, Y., & Han, D. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3260563 · Cited by 36
“Moreover, anomalies are often rare [4], [5], [6].”
BackgroundLin, C., Du, B., Sun, L., & Li, L. (2024) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2024.3360640 · Cited by 39
“Audibert et al. [4] proposed USAD, based on adversarial auto-encoders that isolate anomalies and provide fast training.”
BackgroundAnomaly Detection for Telemetry Time Series Using a Denoising Diffusion Probabilistic Model
Sui, J., Yu, J., Song, Y., & Zhang, J. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3383416 · Cited by 36
“6) USAD [62]: USAD uses adversarial training to construct an AE-based model and detects anomalies through reconstruction errors.”
MethodEnsemble Denoising Autoencoders Based on Broad Learning System for Time-Series Anomaly Detection
Lin, Y., Yu, Z., Yang, K., & Chen, C. L. P. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3548941 · Cited by 30
“Therefore, we used the best-F 1 score to evaluate the model performance, referring to previous works in [33], [35], [40], and [52].”
Influential"Do you know existing accuracy metrics overrate time-series anomaly detections?"
Hwang, W., Yun, J.-H., Kim, J., & Min, B. G. (2022) · ACM Symposium on Applied Computing · doi:10.1145/3477314.3507024 · Cited by 28
“The pointadjust metric [25] is a frequently employed metric in the literature [1, 18, 23].”
AERF: Adaptive ensemble random fuzzy algorithm for anomaly detection in cloud computing
Jiang, J., Liu, F., Ng, W. W. Y., Tang, Q., Zhong, G., Tang, X., & Wang, B. (2023) · Computer Communications · doi:10.1016/j.comcom.2023.01.004 · Cited by 28
Zhong, G., Liu, F., Jiang, J., Wang, B., & Chen, C. L. P. (2023) · Information Sciences · doi:10.1016/j.ins.2023.119914 · Cited by 28
A novel anomaly detection approach based on ensemble semi-supervised active learning (ADESSA)
Niu, Z., Guo, W., Xue, J., Wang, Y., Kong, Z., & Huang, L. (2023) · Computers & Security · doi:10.1016/j.cose.2023.103190 · Cited by 27
Gringotts: Fast and Accurate Internal Denial-of-Wallet Detection for Serverless Computing
Shen, J., Zhang, H., Geng, Y., Li, J., Wang, J., & Xu, M. (2022) · ACM SIGSAC Conference on Computer and Communications Security (CCS) · doi:10.1145/3548606.3560629 · Cited by 38
“USAD [4] inferred system abnormal behavior by combining generative adversarial networks (GANs) and an auto-encoder.”
BackgroundGlobal–Local Association Discrepancy for Multivariate Time Series Anomaly Detection in IIoT
Zhou, X., Dai, C., Wang, W., & Qiu, T. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3330696 · Cited by 27
“USAD [22] proposes an AE-based unsupervised anomaly detection approach for MTS.”
MethodMTAD: Multiobjective Transformer Network for Unsupervised Multisensor Anomaly Detection
Belay, M. A., Rasheed, A., & Rossi, P. S. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3396690 · Cited by 34
“We perform downsampling using a median filter with a 1-min window size and no overlap, as in [34], for both training and test data.”
BackgroundMethodMETER: A Dynamic Concept Adaptation Framework for Online Anomaly Detection
Zhu, J., Cai, S., Deng, F. Y., Ooi, B. C., & Zhang, W. (2023) · Proceedings of the VLDB Endowment · doi:10.14778/3636218.3636233 · Cited by 40
Self-supervised multi-transformation learning for time series anomaly detection
Han, H., Fan, H., Huang, X., & Han, C. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.124339 · Cited by 31
Unsupervised anomaly detection of multivariate time series based on multi-standard fusion
Tian, H., Kong, H., Lu, S., & Li, K. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128634 · Cited by 26
Federated Variational Learning for Anomaly Detection in Multivariate Time Series
Zhang, K., Jiang, Y., Seversky, L. M., Xu, C., Liu, D., & Song, H. (2021) · IEEE International Performance, Computing, and Communications Conference · doi:10.1109/ipccc51483.2021.9679367 · Cited by 44
“Besides, we utilize the point-adjust method to evaluate the performance of models based on a contiguous-segment detecting strategy that is widely adopted in [3], [8], [14], [16], [17].”
BackgroundMethodA novel anomaly detection method for multimodal WSN data flow via a dynamic graph neural network
Zhang, Q., Ye, M., & Deng, X. (2022) · Connection Science · doi:10.1080/09540091.2022.2078281 · Cited by 27
“The authors in (Audibert et al., 2020) designed a two-stage adversarial training-based anomaly detection model.”
BackgroundAnomaly and change point detection for time series with concept drift
Liu, J., Yang, D., Zhang, K., Gao, H., & Li, J. (2023) · World Wide Web · doi:10.1007/s11280-023-01181-z · Cited by 41
“(7)USADisbasedonadverselytrainedautoencoders.”
MethodInfluentialMa, S., Nie, J., Guan, S., He, Z., & Gao, M. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3443910 · Cited by 31
“7) USAD [38]: A reconstruction-based model that combines GANs and AE to enhance detection performance.”
CrossFuN: Multiview Joint Cross-Fusion Network for Time-Series Anomaly Detection
Bai, Y., Wang, J., Zhang, X. H., Miao, X., & Lin, Y. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3315420 · Cited by 31
“We compare CrossFuN with several models that do not consider the intermetric dependencies (MAD-GAN, Omni-Anomaly, and USAD).”
MethodInfluentialHe, S., Li, G., Xie, K., & Sharma, P. K. (2024) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2024.3459631 · Cited by 27
“• USAD [36]: This method trains an encoder-decoder framework in an adversarial manner to achieve fast and efficient training.”
Method6G-XSec: Explainable Edge Security for Emerging OpenRAN Architectures
Wen, H., Sharma, P., Yegneswaran, V., Porras, P. A., Gehani, A., & Lin, Z. (2024) · ACM Workshop on Hot Topics in Networks (HotNets) · doi:10.1145/3696348.3696881 · Cited by 27
Qian, C., Tang, W., & Wang, Y. (2025) · Future Generation Computer Systems · doi:10.1016/j.future.2025.107751 · Cited by 32
SimAD: A Simple Dissimilarity-Based Approach for Time-Series Anomaly Detection
Zhong, Z., Yu, Z., Xi, X., Xu, Y., Cao, W., Yang, Y., Yang, K., & You, J. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3590220 · Cited by 50
“…machine learning approaches into deep networks, enhancing their ability to handle complex data; 2) reconstruction-based approaches [14], [19], [29] involve training models using normal data and leveraging reconstruction error as an anomaly score, attributing higher scores to anomalous…”
InfluentialAslam, M. M., Tufail, A., Gul, H., Irshad, M. N., & Namoun, A. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-025-11320-9 · Cited by 41
Self-adversarial variational autoencoder with spectral residual for time series anomaly detection
Liu, Y., Lin, Y., Xiao, Q., Hu, G., & Wang, J. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.06.030 · Cited by 31
Pham, T.-A., Lee, J., & Park, C.-S. (2022) · Applied Sciences · doi:10.3390/app121910078 · Cited by 30
Zhang, X., Shi, S., Sun, H., Chen, D., Wang, G., & Wu, K. J. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.12.023 · Cited by 28
Xie, T., Xu, Q., Jiang, C., Gao, Z., & Wang, X. Z. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3381790 · Cited by 25
Du, Y., Huang, Y., Wan, G., & He, P. (2022) · Mathematics · doi:10.3390/math10224373 · Cited by 24
Unmasking overestimation: a re-evaluation of deep anomaly detection in spacecraft telemetry
Herrmann, L., Bieber, M., Verhagen, W. J. C., Cosson, F., & Santos, B. F. (2024) · CEAS Space Journal · doi:10.1007/s12567-023-00529-5 · Cited by 31
“…OmniAnomaly [26] 2019 RNN, VAE Reconstruction error 88,57 89,89 84,34 THOC [17] 2020 RNN, One-class network Dissimilarity 93,67 95,18 88,09 USAD [31] 2020 Autoencoder, adverse training Reconstruction error 93,82 91,09 81,86 84,6 GTA [32] 2021 Transformer, GNN Prediction error 91,11 90,04 91…”
DyEdgeGAT: Dynamic Edge via Graph Attention for Early Fault Detection in IIoT Systems
Zhao, M., & Fink, O. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3381002 · Cited by 32
Multivariate Time Series Anomaly Detection via Low-Rank and Sparse Decomposition
Belay, M. A., Rasheed, A., & Rossi, P. S. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3452318 · Cited by 32
Gao, R., Chen, Z., Wu, X., Yu, Y., & Zhang, L. (2024) · Cluster Computing · doi:10.1007/s10586-024-04707-w · Cited by 31
“[36] propose an anomaly detection model based on an adversely trained autoencoder, which can isolate anomalies while providing rapid training and possessing excellent robustness.”
CollaborEM: A Self-Supervised Entity Matching Framework Using Multi-Features Collaboration
Ge, C., Wang, P., Chen, L., Liu, X., Zheng, B., & Gao, Y. (2021) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2021.3134806 · Cited by 50
“Many studies [2], [53] focus on anomaly detection based on the information provided by a single dataset.”
BackgroundResNet-AE for Radar Signal Anomaly Detection
Cheng, D., Fan, Y., Fang, S., Wang, M., & Liu, H. (2022) · Sensors · doi:10.3390/s22166249 · Cited by 25
Precursor-of-Anomaly Detection for Irregular Time Series
Jhin, S. Y., Lee, J., & Park, N. (2023) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3580305.3599469 · Cited by 45
Lee, J., Park, B., & Chae, D. (2023) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3583780.3614857 · Cited by 38
Hybrid graph transformer networks for multivariate time series anomaly detection
Gao, R., He, W. P., Yan, L., Liu, D., Yu, Y., & Ye, Z. (2023) · The Journal of Supercomputing · doi:10.1007/s11227-023-05503-w · Cited by 24
A context-aware unsupervised predictive maintenance solution for fleet management
Giannoulidis, A., & Gounaris, A. (2022) · Journal of Intelligent Information Systems · doi:10.1007/s10844-022-00744-2 · Cited by 33
“The combination of these models with adversarial and meta-learning techniques has shown tremendous results in anomaly detection tasks (Audibert et al., 2020; Tuli et al., 2022).”
BackgroundDUMA: Dual Mask for Multivariate Time Series Anomaly Detection
Pan, J., Ji, W., Zhong, B., Wang, P., Wang, X., & Chen, J. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3225338 · Cited by 25
“For a fair comparison, our approach is consistent with these papers [5], [21], [34], enumerating all possible anomaly thresholds to search for the best F1 score as the primary evaluation metric.”
MethodProtocol study and anomaly detection for server-driven traffic in SCADA networks
Lin, C., & Nadjm‐Tehrani, S. (2023) · International Journal of Critical Infrastructure Protection · doi:10.1016/j.ijcip.2023.100612 · Cited by 22
TPAD: Temporal-Pattern-Based Neural Network Model for Anomaly Detection in Multivariate Time Series
Ma, S., Guan, S., He, Z., Nie, J., & Gao, M. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3327138 · Cited by 23
“Audibert et al. [46] proposed USAD based on adversarial trained autoencoders.”
BackgroundMethodDRAGON: Decentralized Fault Tolerance in Edge Federations
Tuli, S., Casale, G., & Jennings, N. R. (2023) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2022.3199886 · Cited by 21
“Another recent class of methods applies neural networks to reconstruct the last state of the system [18], [29].”
BackgroundInfluentialGuan, S., He, Z., Ma, S., & Gao, M. (2024) · Computers & Security · doi:10.1016/j.cose.2024.103877 · Cited by 27
Variate Associated Domain Adaptation for Unsupervised Multivariate Time Series Anomaly Detection
He, Y., Bian, Y., Ding, X. L., Wu, B., Guan, J., Zhang, J., & Zhou, S. (2024) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3663573 · Cited by 21
“Our algorithm is more robust, achieving the best performance on dataset SWaT → WADI and relatively good AUC performance In most cases, baselines can benefit from one of the DA approaches.”
InfluentialPATE: Proximity-Aware Time Series Anomaly Evaluation
Ghorbani, R., Reinders, M. J. T., & Tax, D. M. J. (2024) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3637528.3671971 · Cited by 29
ART: A Unified Unsupervised Framework for Incident Management in Microservice Systems
Sun, Y., Shi, B., Mao, M., Ma, M., Xia, S., Zhang, S., & Pei, D. (2024) · IEEE/ACM International Conference on Automated Software Engineering (ASE) · doi:10.1145/3691620.3695495 · Cited by 71
Xie, Y., Zhang, H., & Babar, M. A. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3696410.3714941 · Cited by 50
Wang, H., Nie, J., He, Z., Gao, M., Song, W., & Dong, Z. (2023) · Energy Reports · doi:10.1016/j.egyr.2023.01.092 · Cited by 24
An Efficient Framework for Unsupervised Anomaly Detection over Edge-Assisted Internet of Things
Liu, Y., Wang, H., Zheng, X., & Tian, L. (2023) · ACM Transactions on Sensor Networks · doi:10.1145/3587935 · Cited by 26
“[4] proposed a light-weighted method for correlation anomaly detection based on an encoding-decoding mode.”
BackgroundMethodInfluentialM3GAN: A masking strategy with a mutable filter for multidimensional anomaly detection
Li, Y., Peng, X., Wu, Z., Yang, F., He, X., & Li, Z. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.110585 · Cited by 23
Chen, K., Liu, J., Guo, W., & Wang, X. (2023) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2023.108745 · Cited by 26
PAFormer: Anomaly Detection of Time Series With Parallel-Attention Transformer
Bai, N., Wang, X., Han, R., Wang, Q., & Liu, Z. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3337876 · Cited by 25
“USAD [31] consists of an encoder and a decoder, and amplifies the reconstruction error containing abnormal inputs through adversarial training, detecting anomalies in the data, while obtaining stability based on generative adversarial networks (GANs) [32] architecture.”
InfluentialChen, J., Pi, D., & Wang, X. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05395-0 · Cited by 23
He, S., Li, G., Yi, T., Alfarraj, O., Tolba, A. K. R., Sangaiah, A. K., & Sherratt, R. S. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3409391 · Cited by 23
“Existing MTSAD techniques [4], [7], [8], [9], [10], [11] focus primarily on identifying the timestamps where system errors, attacks, and behavior deviations occur.”
TimeSeriesBench: An Industrial-Grade Benchmark for Time Series Anomaly Detection Models
Si, H., Li, J., Pei, C., Cui, H., Yang, J., Sun, Y., Zhang, S., Li, J., Zhang, H., Han, J., et al. (2024) · Proceedings/Proceedings - International Symposium on Software Reliability Engineering · doi:10.1109/issre62328.2024.00017 · Cited by 58
Trustworthy AI-based Performance Diagnosis Systems for Cloud Applications: A Review
Xin, R., Wang, J., Chen, P., & Zhao, Z. (2025) · ACM Computing Surveys · doi:10.1145/3701740 · Cited by 30
Deep unsupervised anomaly detection in high-frequency markets
Poutré, C., Chételat, D., & Morales, M. (2024) · The Journal of Finance and Data Science · doi:10.1016/j.jfds.2024.100129 · Cited by 34
Pre-training Enhanced Transformer for multivariate time series anomaly detection
Wang, C., Shi, H., Hu, J., Yang, X. M., Zhang, J., Du, S., & Li, T. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103171 · Cited by 22
Task-Oriented Self-supervised Learning for Anomaly Detection in Electroencephalography
Zheng, Y., Liu, Z., Mo, R., Chen, Z., Zheng, W., & Wang, R. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-16452-1_19 · Cited by 20
Log Anomaly Detection by Adversarial Autoencoders With Graph Feature Fusion
Xie, Y., & Yang, K. (2023) · IEEE Transactions on Reliability · doi:10.1109/tr.2023.3305376 · Cited by 26
“To address this issue, we draw inspiration from the concept of generative adversarial nets introduced in [52].”
BackgroundUnsupervised anomaly detection by densely contrastive learning for time series data
Zhu, W., Li, W., Dorsey, E. R., & Luo, J. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.09.038 · Cited by 25
Prodigy: Towards Unsupervised Anomaly Detection in Production HPC Systems
Aksar, B., Sencan, E., Schwaller, B., Aaziz, O., Leung, V. J., Brandt, J. M., Kulis, B., Egele, M., & Coskun, A. K. (2023) · International Conference for High Performance Computing, Networking, Storage and Analysis (SC) · doi:10.1145/3581784.3607076 · Cited by 27
A graph attention network-based model for anomaly detection in multivariate time series
Zhāng, W., He, P., Qin, C., Yang, F., & Liu, Y. (2023) · The Journal of Supercomputing · doi:10.1007/s11227-023-05772-5 · Cited by 21
“Another study [7] adopted the architecture of AE, consisting of one encoder and two decoders, and leveraged an adversarial learning technique to train the model.”
Unraveling the ‘Anomaly’ in Time Series Anomaly Detection: A Self-supervised Tri-domain Solution
Sun, Y., Pang, G., Ye, G., Chen, T., Hu, X., & Yin, H. (2024) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde60146.2024.00080 · Cited by 48
Gao, R., Wang, J., Yu, Y., Wu, J., & Zhang, L. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.129168 · Cited by 23
Gao, C., Ma, H., Pei, Q., & Chen, Y. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06412-6 · Cited by 20
“In this section, we compare the performance of the D-GATAD model with eight popular multivariate time series (MTS) anomaly detection methods, namely MAD-GAN [36], OmniAnomaly [37], USAD [38], TranAD [39], BTAD [40], GST-Pro [41], ESTAD [42] and SP-GATs [25].”
Miao, Q., Xu, C., Zhan, J., Zhu, D., & Wu, C. (2023) · Communications in computer and information science · doi:10.1007/978-981-99-1645-0_42 · Cited by 27
KalmanAE: Deep Embedding Optimized Kalman Filter for Time Series Anomaly Detection
Huang, X., Zhang, F., Wang, R., Lin, X., Liu, H., & Fan, H. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3329098 · Cited by 25
“USAD [12] proposes a GAN-based anomaly detection algorithm for multivariate time series, which uses two autoencoders trained in stages, and although it improves the shortcomings of GAN, it does not have a targeted design for multivariate time series modeling itself.”
BackgroundMethodXu, Q., Xie, T., Jiang, C., Cheng, Q., & Wang, X. Z. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3413952 · Cited by 19
“EncDec-AD[19]is anLSTM-basedautoencoder,whileUSAD[29]integratesthe generativeadversarialstrategyintoatraditionalautoencoder.”
MethodKea, K., Kim, D., Huot, C., Kim, T.-K., & Han, Y. (2024) · Entropy · doi:10.3390/e26110954 · Cited by 36
MAFCD: Multi-level and adaptive conditional diffusion model for anomaly detection
Wu, Z., Zhu, L., Yin, Z., Xu, X., Jianmin, Z., Wei, X., & Yang, X. S. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.102965 · Cited by 24
Can Industrial Intrusion Detection Be SIMPLE?
Wolsing, K., Thiemt, L., van Sloun, C., Wagner, E., Wehrle, K., & Henze, M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-17143-7_28 · Cited by 20
“(4 4 ) Autoencoder (15) [12, 16, 26, 36, 40, 41, 47, 52, 58, 62, 68, 75, 77, 83] [12, 62] [45]”
Time Series Anomaly Detection via Reinforcement Learning-Based Model Selection
Zhang, J. E., Wu, D., & Boulet, B. (2022) · Conference proceedings - Canadian Conference on Electrical and Computer Engineering · doi:10.1109/ccece49351.2022.9918216 · Cited by 22
“For USAD, we use the implementation from the authors’ original GitHub repository 2.”
MethodInfluentialLuo, X., Jiang, Y., Wang, E., & Men, X. (2022) · EURASIP Journal on Advances in Signal Processing · doi:10.1186/s13634-022-00943-7 · Cited by 21
Xu, L., Ding, X., Zhao, D., Liu, A. X., & Zhang, Z. (2023) · Entropy · doi:10.3390/e25020180 · Cited by 22
“USAD combines generative adversarial networks (GAN) and autoencoders to model multidimensional time series.”
InfluentialBaek, M. H., & Kim, S. B. (2023) · IEEE Access · doi:10.1109/access.2023.3281407 · Cited by 26
“Furthermore, when the correlations between variables were taken into account using the signature matrix, the performances of the LOF, OCSVM, iForest, AE, VAE and USAD models were improved.”
BackgroundMethodInfluentialHybridAD: A Hybrid Model-Driven Anomaly Detection Approach for Multivariate Time Series
Lin, W., Wang, S., Wu, W., Li, D., & Zomaya, A. Y. (2023) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2023.3290027 · Cited by 23
“In this regard, unsupervised methods are frequently adopted as a promising alternative to supervised or semi-supervised methods for anomaly detection [7], [8], [9], [10].”
BackgroundMethodInfluentialKhan, M. A., Iqbal, N., Jamil, H., Qayyum, F., Jang, J., Khan, S., Kim, J., & Kim, D. (2023) · Future Generation Computer Systems · doi:10.1016/j.future.2023.11.036 · Cited by 22
Archetype analysis: A new subspace outlier detection approach
Fayos, I. C., Epifanio, I., Piérola, A., & Ballester, A. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.106830 · Cited by 18
Yang, Q., Zhang, J., Zhang, J., Sun, C., Xie, S., Liu, S., & Ji, Y. (2024) · Electronics · doi:10.3390/electronics13112032 · Cited by 18
Aslam, M. M., Silva, L. C. D., Apong, R. A. A. H. M., & Tufail, A. (2025) · Scientific Reports · doi:10.1038/s41598-025-12775-0 · Cited by 30
Anomaly Detection and Localization in NFV Systems: an Unsupervised Learning Approach
Johari, S. S., Shahriar, N., Tornatore, M., Boutaba, R., & Saleh, A. (2022) · IEEE/IFIP Network Operations and Management Symposium (NOMS) · doi:10.1109/noms54207.2022.9789938 · Cited by 18
“Autoencoder is another popular UL method that generally trains faster and requires much less training data than RNNs to reach a satisfactory performance [8].”
BackgroundMethodInfluentialDetecting Extreme Traffic Events Via a Context Augmented Graph Autoencoder
Hu, Y., Qu, A., & Work, D. B. (2022) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3539735 · Cited by 19
Disentangled Dynamic Deviation Transformer Networks for Multivariate Time Series Anomaly Detection
Wang, C., Xing, S., Gao, R., Yan, L., Xiong, N. N., & Wang, R. (2023) · Sensors · doi:10.3390/s23031104 · Cited by 20
“Audibert et al. [24] proposed a multivariate time series unsupervised approach based on back-trained autoencoders, which has greater generalization capabilities, by combining the advantages of autoencoders and adversarial training.”
BackgroundMethodFedTADBench: Federated Time-series Anomaly Detection Benchmark
Liu, F., Zeng, C. H., Zhang, L., Zhou, Y., Mu, Q., Zhang, Y., Zhang, L., & Zhu, C. (2022) · IEEE International Conference on High Performance Computing and Communications; IEEE International Conference on Smart City; IEEE International Conference on Data Science and Systems (HPCC/SmartCity/DSS) · doi:10.1109/hpcc-dss-smartcity-dependsys57074.2022.00071 · Cited by 20
Mutual information maximization for semi-supervised anomaly detection
Liu, S., & Tian, M. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.111196 · Cited by 18
Zhu, K. N., Song, P., & Zhao, C. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3371109 · Cited by 17
“To demonstrate our superiority, we selected 14 anomaly detection models as comparison, including IF [4], PCA [18], SFA [20], CVA [21], DAGMM [27], GDN [26], SG-AE [28], OmniAnomaly [24], USAD [29], AAU-Net [25], TranAD [12], TopoMAD [32], MSCRED [30], and MTAD-GAT [31].”
Lee, B., Kim, S., Maqsood, M., Moon, J., & Rho, S. (2024) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2024.054826 · Cited by 30
RESTAD: Reconstruction and Similarity Based Transformer for Time Series Anomaly Detection
Ghorbani, R., Reinders, M. J. T., & Tax, D. M. J. (2024) · Machine learning for signal processing ... · doi:10.1109/mlsp58920.2024.10734755 · Cited by 20
“The underlying assumption is that typical data will have lower reconstruction errors, whereas anomalous data will exhibit higher errors due to the unfamiliarity of the model with these patterns [10, 12].”
BackgroundMethodWang, S., Xu, Q., Zhu, S., & Wang, B. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.125736 · Cited by 24
Robust System Instance Clustering for Large-Scale Web Services
Zhang, S., Li, D., Zhong, Z., Zhu, J., Liang, M., Luo, J., Sun, Y., Su, Y., Xia, S., Hu, Z., et al. (2022) · International World Wide Web Conference (WWW) · doi:10.1145/3485447.3511983 · Cited by 18
Merlion: A Machine Learning Library for Time SeriesPreprint
Bhatnagar, A., Kassianik, P., Liu, C., Lan, T., Yang, W., Cassius, R., Sahoo, D., Arpit, D., Subramanian, S., Woo, G., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2109.09265 · Cited by 57
“Time series are ubiquitous in monitoring the behavior of complex systems in real-world applications, such as IT operations management, manufacturing industry and cyber security (Hundman et al., 2018; Mathur and Tippenhauer, 2016; Audibert et al., 2020).”
BackgroundUnsupervised Contextual Anomaly Detection for Database Systems
Li, S., Yin, Q., Li, G., Li, Q. L., Liu, Z., & Zhu, J. (2022) · ACM SIGMOD International Conference on Management of Data (SIGMOD) · doi:10.1145/3514221.3517861 · Cited by 23
“The only one exception is the FNR on A 1 of Scenario-I (about 0.19), which is higher than that of USAD [11] (about 0.09).”
BackgroundMethodResultInfluentialAnomaly detection in the context of long-term cloud resource usage planning
Nawrocki, P., & Sus, W. (2022) · Knowledge and Information Systems · doi:10.1007/s10115-022-01721-5 · Cited by 20
Sparse Non-Linear Vector Autoregressive Networks for Multivariate Time Series Anomaly Detection
Belay, M. A., Rasheed, A., & Rossi, P. S. (2024) · IEEE Signal Processing Letters · doi:10.1109/lsp.2024.3520019 · Cited by 20
Metadata
- Authors
- Audibert, J., Michiardi, P., Guyard, F., Marti, S., & Zuluaga, M. A.
- Affiliations
- EURECOM (FR)
- EURECOM, Biot, France
- Orange (France) (FR)
- Venue
- Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining
- Publisher
- ACM
- Published
- 2020-08-20
- Pages
- 3395-3404
- OpenAlex
- W3081497074
- Cited by
- 1,094
- References
- 20
- Citations by year
- 2026: 281 · 2025: 266 · 2024: 256 · 2023: 170 · 2022: 84 · 2021: 37
- FWCI
- 27.79
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) (publishedVersion, cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Anomaly Detection Techniques and Applications
- Physical Sciences › Computer Science › Computer Networks and Communications › Software System Performance and Reliability
- Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
- Keywords
- unsupervised anomaly detection, multivariate time series, adversarial autoencoder, network anomaly detection, orange, scalability, robustness
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.