USAD: UnSupervised Anomaly Detection on Multivariate Time Series
Open accessThe DOI matches its Crossref record; 19 of the 20 references it cites were found in a registry or an index.
Abstract
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 401–500 of 1,000
Novikova, E. S., Владимировна, Ф. Е., Danilov, A., & Saenko, I. B. (2024) · SN Computer Science · doi:10.1007/s42979-024-02704-9 · Cited by 6
“Audibert et al. [13] investigated the applicability of autoencoders for anomaly detection in the unsupervised learning mode.”
ALOC: Attack-Aware by Utilizing the Adversarially Learned One-Class Classifier for SCADA System
Li, W., Yao, Y. K., Sheng, C., Zhang, N., & Yang, W. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3384437 · Cited by 6
“The detection process maps to lines 13–23, the anomaly score is calculated by the sum of both the reconstruction error and the discrimination error, the parameters λ G and λ D is to balance the importance between G and D for detecting in line 18, and we set them by [41].”
MethodEA-GAT: Event aware graph attention network on cyber-physical systems
Yağcı, M. Y., & Aydın, M. A. (2024) · Computers in Industry · doi:10.1016/j.compind.2024.104097 · Cited by 7
An IIoT Temporal Data Anomaly Detection Method Combining Transformer and Adversarial Training
Tian, Y., Wang, W., & He, J. (2024) · International Journal of Information Security and Privacy · doi:10.4018/ijisp.343306 · Cited by 5
Tafazoli, S., Lu, Y., Wu, R., Srinivas, T. V. A., Cruz, H. D., Mercer, R., & Keogh, E. (2024) · Knowledge and Information Systems · doi:10.1007/s10115-024-02107-5 · Cited by 8
“For example, “ These two subsequences are alike ” (motifs [54]) or “ this subsequence is different to all others ” (discords) [4–6, 39 ], whereas C22 simply However, there are two difficulties in attempting to unify these two ideas.”
Lachure, J., & Doriya, R. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3472022 · Cited by 10
CATS: Contrastive learning for Anomaly detection in Time Series
Ky, J. R., Mathieu, B., Lahmadi, A., & Boutaba, R. (2024) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata62323.2024.10825476 · Cited by 11
“These can be categorized into five groups: reconstruction-based [1]–[4], distance-based [5], one-class classification-based [6], [7], isolation-based [8], and generative-based [9], [10].”
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
Wang, Y., Sun, H., Wang, C., Zhu, M., Wang, J., Tang, W., Qi, Q., Zhuang, Z., & Liao, J. (2024) · Proceedings · doi:10.1109/icdm59182.2024.00107 · Cited by 16
“Some focus on spatial characteristics, neglecting temporal dependencies [5], while others, like [6], [10], capture temporal patterns.”
Assessing Unsupervised Machine Learning solutions for Anomaly Detection in Cloud Gaming Sessions
Ky, J. R., Mathieu, B., Lahmadi, A., & Boutaba, R. (2022) · International Conference on Network and Service Management (CNSM) · doi:10.23919/cnsm55787.2022.9964533 · Cited by 8
“Previous works such as [16], [24], tested different thresholds and reported those that led to the best F1-score which explains why they got better results.”
ResultA Dilated Transformer Network for Time Series Anomaly Detection
Wu, B., Yao, Z., Tu, Y., & Chen, Y. (2022) · IEEE International Conference on Tools with Artificial Intelligence (ICTAI) · doi:10.1109/ictai56018.2022.00016 · Cited by 5
“We compare our model with 7 state-of-the-art baselines, including RNN-based methods: LSTM-VAE [11], OmniAnoma-ly [17]; Graph-based method: MTAD-GAT [21]; CNN-based method: USAD [22]; Transformer-based methods: TransFram [23], AnomalyTran [24] and TranAD [25].”
Lundström, A., O’Nils, M., & Qureshi, F. Z. (2023) · IEEE Access · doi:10.1109/access.2023.3310653 · Cited by 7
“In the evaluation, five different types of benchmark datasets for multivariate anomaly detection are used that have been widely applied in previous studies [5], [6], [7], [10], [13],”
MethodVariational Autoencoders (VAEs) for Anomaly Detection
Tahir, S. (2024) · Advances in Digital Crime, Forensics, and Cyber Terrorism · doi:10.4018/979-8-3693-8944-7.ch009 · Cited by 4
Detecting cyberattacks based on deep neural network approaches in industrial control systems
Ayas, S., Ayas, M. S., Cavdar, B., & Sahin, A. K. (2025) · Journal of Information Security and Applications · doi:10.1016/j.jisa.2025.104206 · Cited by 4
A Dual-Discriminator Generative Adversarial Network for Anomaly Detection
Ding, D., Wang, Y., Tao, H., Wu, J. X., & Cao, J. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3585978 · Cited by 4
“For comparison, we select seven state-of-the-art unsupervised anomaly detection methods: USAD [39] (adversarial AEs), OmniAnomaly [35] (stochastic RNN), TranAD [40] (transformer-based), GDN [41] (graph neural network), BeatGAN [9] (beat-aware GAN), FGANomaly [10] (pseudo-label enhanced GANomaly),…”
Bose, R., Nwoye, C. I., Lazo, J. F., Lavanchy, J. L., & Padoy, N. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-05141-7_18 · Cited by 4
Xu, Y., Fan, R.-Z., Luo, Y., Ke, W., Zhang, Y., & Zhang, M. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3625242 · Cited by 5
“USAD [7] further enhanced detection performance by coupling autoencoders with adversarial training to amplify reconstruction errors for anomalous samples.”
Naqvi, S. M. R., Ghufran, M., Varnier, C., Nicod, J., & Noureddine Zerhouni, A. (2026) · Journal of Process Control · doi:10.1016/j.jprocont.2026.103677 · Cited by 6
Antinoise Adaptive Time–Frequency Fusion for multivariate time series anomaly detection
Huang, S., Xi, L., Huang, X., Cheng, Y., & Liu, H. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.114584 · Cited by 7
HYDRA: A Multi-Level Hierarchy-Driven Approach for Robust Anomaly Detection in Time Series
Huang, M., Liu, Q., Boniol, P., & Paparrizos, J. (2026) · Proceedings of the ACM on Management of Data · doi:10.1145/3802074 · Cited by 7
“Univariate baselines: KMeansAD [114], Spectral Residual (SR) [97], NormA [13], Sub-KNN [96], Matrix Profile (MP/STOMP) [119], Se-ries2Graph [15], POLY [53], Sub-PCA [2], USAD [7], CNN [70], LSTM-AD [66], MOMENT (fine-tuned; MOMENT-FT) [34], and TimesFM [25].”
The Power of Anomaly Detection in Predictive Maintenance: [Experiments & Analysis]
Papadopoulos, A., Giannoulidis, A., Gounaris, A., & Paparrizos, J. (2026) · Proceedings of the ACM on Management of Data · doi:10.1145/3802119 · Cited by 5
Nweke, L. O. (2026) · Discover Computing · doi:10.1007/s10791-026-10343-2 · Cited by 4
He, J., Dong, Z., & Huang, Y. (2024) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls61622.2024.10606841 · Cited by 5
“These models then utilize the disparity between the predicted and actual data as an anomaly score, labeling timestamps with elevated anomaly scores as outliers [5, 6].”
MethodDiffusion Model in Normal Gathering Latent Space for Time Series Anomaly Detection
Han, J., Feng, S., Zhou, M. Q., Zhang, X., Ong, Y.-S., & Li, X. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-70352-2_17 · Cited by 12
Gmad: multivariate time series anomaly detection based on graph matching learning
Kong, J., Wang, K., Jiang, M., & Tao, X. (2024) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-024-02482-z · Cited by 4
“• USAD: USAD [44] adopts an unsupervised learning approach using an adversarially trained architecture.”
Anomaly Detection Using Graph-Based Autoencoder with Graph Structure Learning Layer
Ahmad, A., Kovalenko, A., & Makarov, I. (2024) · IEEE International Symposium on Logistics and Industrial Informatics (LINDI) · doi:10.1109/lindi63813.2024.10820392 · Cited by 6
Towards Unbiased Evaluation of Time-series Anomaly Detector
Bhattacharya, D., Mukherjee, S., Kamanchi, C., Ekambaram, V., Jati, A., & Dayama, P. S. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890568 · Cited by 4
“• Many studies still use F 1 PA despite its limitations [9, 10, 11, 12, 13, 14], even for evaluating state-of-the-art models [15, 16].”
Generality-aware self-supervised transformer for multivariate time series anomaly detection
Cho, Y., Lee, J.-H., Ham, G., Jang, D., & Kim, D. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06481-7 · Cited by 4
“We employ a well-known technique in this field called replication padding [9, 13, 23], which was introduced in [23], to provide each data point with its local context.”
InfluentialDGAN: Flight Data Anomaly Detection Based on Dual-View Graph Attention Network
Yan, W., Li, X., Zheng, L., Shang, J., Wu, L., Liu, H., Li, X., & Qian, Y. X. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3562842 · Cited by 4
“…time series (MTS) data; recent advancements in deep learning for MTS anomaly detection have introduced methods such as autoencoders [31], [32], transformer architectures [17], [25], [33], [34], [35], generative adversarial networks (GANs) [36], [37], and graph neural networks (GNNs)…”
Transformer-GAN architecture for anomaly detection in multivariate time series
蔡, 美., 汪, 家., 刘, 金., 唐, 朝., & 谢, 永. (2022) · Scientia Sinica Informationis · doi:10.1360/ssi-2022-0133 · Cited by 4
Unsupervised Anomaly Detection Approach for Shift Quality Assessment Using Deep Neural Networks
Oh, G., Park, J., Hwang, K., & Lim, S. (2022) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv51971.2022.9827200 · Cited by 4
Ming, L., Han, D., & Li, D. (2023) · International Database Engineering and Applications Symposium (IDEAS) · doi:10.1145/3589462.3589489 · Cited by 4
Zhang, W., Zhang, S., Chen, J., & Huang, H. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191810 · Cited by 4
“IF: Isolated Forest (IF) [28] builds ensembles of isolation trees to separate fraud samples from normal samples.”
BackgroundMethodFast Anomaly Detection For Multivariate Industrial Time Series Data
Kumar, S., Gupta, G. R., & Parihar, R. P. S. (2023) · International Conference on Computing, Communication and Networking Technologies (ICCCNT) · doi:10.1109/icccnt56998.2023.10307775 · Cited by 4
“[11, 12].”
Enhanced KPI Anomaly Detection: An Unsupervised Hybrid Model with Dynamic Threshold
Wang, Y., Chen, T., Tang, Y., & Huang, L. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10446434 · Cited by 4
Wang, R., & Li, J. (2024) · Information Technology And Control · doi:10.5755/j01.itc.53.1.34254 · Cited by 4
Effective Anomaly Detection for Microservice Systems with Real-Time Feature Selection
Zhou, S., Zhan, X., Li, L., & Liu, Y. (2023) · Asia-Pacific Software Engineering Conference (APSEC) · doi:10.1109/apsec60848.2023.00020 · Cited by 4
Unsupervised Multi-Head Attention Autoencoder for Multivariate Time-Series Anomaly Detection
Kim, M., & Park, S. (2024) · IEEE International Conference on Big Data and Smart Computing (BigComp) · doi:10.1109/bigcomp60711.2024.00011 · Cited by 4
Time-Series to Image-Transformed Adversarial Autoencoder for Anomaly Detection
Kang, J., Kim, M., Park, J., & Park, S. (2024) · IEEE Access · doi:10.1109/access.2024.3450709 · Cited by 7
ALAD: A New Unsupervised Time Series Anomaly Detection Paradigm Based on Activation Learning
Ding, F., Li, B., Ben, X., Zhao, J., & Zhou, H. (2024) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2024.3453762 · Cited by 5
“Forexample,UnsupervisedAnomalyDetection (USAD)[35]adoptsanautoencoderarchitecturewithinanad-versarialtrainingframework,whichamplifiesthereconstruction errorofanomaliesandcontributestoanomalydetection.”
CluSAD: Self-Supervised Learning-Based Anomaly Detection for Industrial Control Systems
Shen, W.-Q., Dai, T., Chen, Z., & Meng, J.-X. (2024) · 2024 5th International Conference on Electronic Communication and Artificial Intelligence (ICECAI) · doi:10.1109/icecai62591.2024.10675256 · Cited by 2
“Audibert et al. [18] proposed a fast and stable anomaly detection method called USAD, which uses a reverse-trained autoencoder to amplify the reconstruction error of anomalous data points in MTS. Unlike existing methods, the CluSAD proposed in this paper overcomes the limitations of traditional…”
MethodZhang, C., Zhang, Y., Peng, L., Wen, Q., Yang, Y., Fan, C., Jiang, M., Fan, L., & Sun, L. (2024) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3627673.3679128 · Cited by 6
Efficient anomaly detection method for offshore wind turbines
Li, Y., Zhang, H., Gao, J.-W., Zhang, Y., Li, P.-F., & Du, H. (2024) · Journal of Electronic Science and Technology · doi:10.1016/j.jnlest.2024.100285 · Cited by 7
Multivariate Time Series Anomaly Detection based on Pre-trained Models with Dual-Attention Mechanism
Sun, Y., Guo, Y., Liang, M., Wen, X., Kuang, J., Zhang, S., Li, H., Xia, K., & Pei, D. (2024) · IEEE International Symposium on Software Reliability Engineering Workshops (ISSREW) · doi:10.1109/issrew63542.2024.00050 · Cited by 5
Scalable and accurate online multivariate anomaly detection
Salles, R. P., Lange, B., Akbarinia, R., Masséglia, F., Ogasawara, E., & Pacitti, E. (2025) · Information Systems · doi:10.1016/j.is.2025.102524 · Cited by 6
“2020 USAD [Audibert et al., 2020, Ang et al., 2023, Lu et al., 2023, Ahmed et al., 2022] DL √ √ Pytorch 2019 AE [Schreyer et al., 2017, Lu et al., 2023, Jacob et al., 2021, Berahmand et al., 2024] DL √ Tensorflow 2019 OmniAnomaly [Su et al., 2019, Ahmed et al., 2022] DL √ Tensorflow 2018 LSTM-VAE…”
InfluentialReal-Time Anomaly Detection for Large-Scale Network Devices
Lei, T., Zhang, S., Kuang, J., Guo, X., & Yang, C. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3529861 · Cited by 5
“MSCRED [10], USAD [11], DOMI [25], OmniAnomaly [7] and InterFusion [13] build anomaly detection models by learning the anomaly patterns using a large span of historical data.”
Detecting and explaining anomalies in industrial Internet of things systems using an autoencoder
Levshun, D., Levshun, D. S., & Kotenko, I. V. (2025) · Ontology of Designing · doi:10.18287/2223-9537-2025-15-1-96-113 · Cited by 7
“АК может быть включѐн в генеративно - состязательную сеть ( Generative Adversarial Network, GAN ) , как это сделано в USAD [19], или объединѐн с графовой НС, как в FuSAGNet [20].”
USD: Unsupervised Soft Contrastive Learning for Fault Detection in Multivariate Time Series
Liu, H., Qiu, X., Shi, Y., & Zang, Z. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890487 · Cited by 11
“[16] 78.8(±1.0) 86.1(±0.9) 78.0(±0.2) 57.0(±0.1) 86.9(±11.7) 18.8(±0.6) 19.8(±0.5) 57.9(±3.6) 31.3(±0.0) 68.4(±19.9) DAGMM [17] 72.8(±3.0) 77.2(±0.9) 64.6(±2.6) 56.5(±2.6) 78.0(±9.2) - - - - - GANF [18] 79.8( Dataset.”
InfluentialA Multi-Source Log Hidden Anomaly Detection Method Integrating Trans-Encoder and LSTM
Luo, C., Jiang, H., Zhou, P., Wang, X., Shao, Z., & Sun, W. (2025) · IEEE Access · doi:10.1109/access.2025.3556959 · Cited by 6
Partially-Supervised Graph Derivation Network With Meta-Learning for Time-Series Anomaly Detection
Zhu, S., Li, Y., Xu, K., & Xu, J. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3558273 · Cited by 4
AIOpsArena: Scenario-Oriented Evaluation and Leaderboard for AIOps Algorithms in Microservices
Sun, Y., Wang, J., Li, Z., Nie, X., Ma, M., Zhang, S., Ji, Y., Zhang, L., Long, W., Chen, H., et al. (2025) · IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER) · doi:10.1109/saner64311.2025.00082 · Cited by 11
Data‐Driven Anomaly Detection and Mitigation for FACTS‐Based Wide‐Area Voltage Control System
Singh, V. K., Govindarasu, M., & Nuqui, R. F. (2025) · IET Cyber-Physical Systems Theory & Applications · doi:10.1049/cps2.70020 · Cited by 6
Sai, B. K., Mayer, Y. T., & Bauernhansl, T. (2021) · Procedia CIRP · doi:10.1016/j.procir.2021.11.061 · Cited by 4
Ferrari, L. M., Hanna, G. A., Volpe, P., Ismailova, E., Brémond, F., & Zuluaga, M. A. (2021) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc46164.2021.9630901 · Cited by 4
“The network is trained using a two-phase adversarial training scheme to allow the AEs to learn how to amplify the reconstruction error as detailed in [2].”
MethodSelf-citationMultiple Temporal Context Embedding Networks for Unsupervised time Series Anomaly Detection
Li, H., Peng, X., Zhuang, H., & Lin, Z. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9747668 · Cited by 6
CNTS: Cooperative Network for Time Series
Yang, J., Shao, Y., & Li, C. (2023) · IEEE Access · doi:10.1109/access.2023.3259467 · Cited by 4
“The Unsupervised Anomaly Detection (USAD) [18] method introduces another decoder based on the traditional autoencoder, and trains the model unsupervised through adversarial learning.”
MethodOswal, S., Shinde, S. K., & Vijayalakshmi, M. (2023) · Communications in computer and information science · doi:10.1007/978-3-031-35644-5_17 · Cited by 5
Hybrid Deep Learning Model for Time Series Anomaly Detection
Han, C. S., & Lee, K. M. (2023) · ACM Conference on Research in Adaptive and Convergent Systems (RACS) · doi:10.1145/3599957.3606232 · Cited by 4
Anomaly Detection in Time Series Satellite Data Using a Deep Learning Method
Han, X., Wang, Y., Zhang, X., & Xu, N. (2023) · CAA Symposium on Fault Detection, Supervision and Safety for Technical Processes (SAFEPROCESS) · doi:10.1109/safeprocess58597.2023.10295895 · Cited by 4
Multivariate Time Series Prediction based on Improved Transformer Model in Computing System
Gong, M., Sun, J., Xie, X., & Zheng, Y. (2023) · International Conference on Cloud Computing, Big Data Application and Software Engineering (CBASE) · doi:10.1109/cbase60015.2023.10439137 · Cited by 4
“The key of Transformer is the self-attention mechanism, and its atomic operation, namely dot-product attention, results in a time complexity and memory usage of O(L 2 ) for each layer[2].”
BackgroundYang, X., Palmes, P., Jha, S., Turkkan, B. O., Vanloo, G., Bagehorn, F., Narayanaswami, C., Shwartz, L., Abe, N., Deng, Y., et al. (2024) · IEEE ... International Conference on Cloud Computing · doi:10.1109/cloud62652.2024.00040 · Cited by 4
“• Evaluation : Existing anomaly detection approaches generally employ F1 score for evaluation [15], [18], [22].”
Lee, M., Lin, J., & Katsikas, S. K. (2024) · Concurrency and Computation Practice and Experience · doi:10.1002/cpe.8288 · Cited by 5
A Prediction-Augmented AutoEncoder for Multivariate Time Series Anomaly Detection
Gong, S., Wu, Z., Liu, Y., Lin, Y., & Wang, J. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-92185-9_56 · Cited by 4
Approach to Anomaly Detection in Cyber-Physical Object Behavior
Shulepov, A., Novikova, E. S., & Murenin, I. (2022) · Studies in computational intelligence · doi:10.1007/978-3-030-96627-0_38 · Cited by 4
Liu, X., Zhi, X., Zhou, T., Zhao, L., Tian, L., Gao, R., Luo, J., Cui, W., & Wang, Q. (2025) · Scientific Data · doi:10.1038/s41597-025-05591-8 · Cited by 5
Wang, S., Liu, S., Fan, X., Wang, H., Zhou, Y., Gao, H., Lu, H., & Deng, X. (2024) · IEEE Smart World Congress (SWC) · doi:10.1109/swc62898.2024.00094 · Cited by 3
“In recent years, significant advancements have been made in deep learning based methods, which can be broadly categorized into reconstruction-based such as LSTM-AE [19] and USAD [20], and prediction-based methods like THOC [21].”
Cui, Z., Zang, D., Zhang, J., & Tang, K. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3577826 · Cited by 6
“5) USAD [36]: Improves anomaly detection performance using AE with adversarial training.”
Frequency-Domain Anomaly Detection for Encrypted Traffic in Industrial Control Systems
Wu, Z., Li, H., Hua, Y., Tran, N. H. K., & Gan, H. (2025) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2025.3582394 · Cited by 6
“Thesemethodsen-compassMLPmodels[10],[11],AEapproaches[12],[13], [14],[15],andCNNframeworks[16],[17].”
InfluentialMultivariate Time Series Anomaly Detection With Hierarchical Component-Aware
Liu, H., Zhang, Z., Xi, L., & Zhang, F. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3605933 · Cited by 3
“USAD [23] introduced an encoder-decoder architecture within an adversarial training framework.”
Time-Series Anomaly Detection for Sensor Data: Models, Metrics, and Methodologies—A Review
Sayyaf, M. I., Pascacio, P., Zhu, N., & Renaudin, V. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3616395 · Cited by 7
“Audibert et al. [4] propose unsupervised anomaly detection (USAD).”
Cui, K., Gao, L., Deng, X., Liu, S., Yi, L., He, S., & Lu, H. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3614179 · Cited by 4
Sun, S., Liu, X., Li, H., Wang, Z., & Zhu, Y. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06927-y · Cited by 3
Xiu, Z., He, Q., Zhong, Y., & Chen, J. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11228289 · Cited by 3
Sensor fusion transformer for interpretable anomaly detection via multi-sensor time series data
Chung, H. T., Chung, J., & Bae, S. J. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.113049 · Cited by 4
Benabderrahmane, S., Cheney, J., & Rahwan, T. (2025) · Scientific Reports · doi:10.1038/s41598-025-25621-0 · Cited by 6
He, S., Guo, Q., Feng, K., Liang, D., Xie, K., & Sharma, P. K. (2025) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2025.3640165 · Cited by 3
Liu, Y., & Carlo, S. D. (2025) · Mechanical Systems and Signal Processing · doi:10.1016/j.ymssp.2025.113767 · Cited by 6
Lee, S. K., Kim, W., Lee, S., Park, J., Chun, S., Yeung, H., & Ko, H. (2026) · Virtual and Physical Prototyping · doi:10.1080/17452759.2025.2611194 · Cited by 4
Fined-grained unmanned aerial vehicle anomaly detection by fusing multi-domain features
Zhou, J., & Zhu, Y. (2026) · Cybersecurity · doi:10.1186/s42400-025-00422-0 · Cited by 13
“In Audibert et al. (2020), researchers failed to consider fined-grained feature error.”
InfluentialTransformer-Driven Multicenter Manifold Modeling Deep SVDD for Anomaly Detection in IIoT Networks
Zhong, W., Li, D., Zou, Y., Guan, T., & Wang, W. W. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3669246 · Cited by 6
“USAD [45] employs an adversarial AE architecture to achieve fast and e ff ective anomaly detection in an unsupervised setting.”
Augmented temporal grid for graph neural networks and anomaly detection in spacecraft
Kiprit, G. N., Koch, A. W., Petry, M. J., & Werner, M. (2026) · Astrodynamics · doi:10.1007/s42064-025-0278-0 · Cited by 4
“Examples of forecasting-based methods include DeepAnT [6], N-BEATS [7], Telemanom [8], Random Forests [9], and Gradient Boosting Decision Trees [10], and examples for reconstruction-based methods are Convolutional Autoencoder [11], USAD [12], OmniAnomaly [13], and RobustPCA [14].”
An Evaluation of Time-Series Anomaly Detection in Computer Networks
Nguyen, H. T., Hajisafi, A., Abdoli, A., Kim, S. H., & Shahabi, C. (2023) · International Conference on Information Networking · doi:10.1109/icoin56518.2023.10049051 · Cited by 6
Deep Learning-Based Anomaly Detection for Time-Series Data in Industrial Control Systems
Zhang, Y. A., Li, B., & Zhang, X. (2023) · Chinese Control Conference (CCC) · doi:10.23919/ccc58697.2023.10241095 · Cited by 3
UDAD: An Accurate Unsupervised Database Anomaly Detection Method
Zhong, H., Zhang, F., Zhao, Y., Zhang, W., Xiao, W., Tang, X., & Zang, L. (2023) · IEEE International Performance, Computing, and Communications Conference · doi:10.1109/ipccc59175.2023.10253824 · Cited by 4
Automated Machine Learning Configuration to Learn Intrusion Detectors on Attack-Free Datasets
Anser, O., François, J., & Chrisment, I. (2024) · Conference on Local Computer Networks · doi:10.1109/lcn60385.2024.10639690 · Cited by 6
Uncompromised Accuracy: Fast and Reliable Multivariate Anomaly Detection for Satellite Signals
Sadr, M. A. M., & Qaraqe, M. (2024) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2024.3463629 · Cited by 4
Yang, B., Ma, T., Rong, H., Huang, X., Wang, Y., Zhao, B., & Wang, C. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06310-x · Cited by 6
“GAN-based approaches, such as the one proposed by Audibert et al. [19], use adversarial training between a generator and discriminator to isolate anomalous data points.”
Sensor Fusion Enhances Anomaly Detection in a Flood Forecasting System
Ma, A., Karande, A., Dahlquist, N., Ferrero, F., & Nguyen, N. R. (2025) · Journal of Sensor and Actuator Networks · doi:10.3390/jsan14020034 · Cited by 5
CBMAD: Anomaly Detection in IoT Network Traffic via Consistent Bidirectional Mamba Autoencoder
Yu, Y.-C., Ouyang, Y., & Lin, C.-A. (2025) · IEEE International Conference on High Performance Switching and Routing (HPSR) · doi:10.1109/hpsr64165.2025.11038914 · Cited by 3
“AVG F-scores USAD [15] 0.8547 TAnoGAN [17] 0.8179 TcnED [13] 0.9156 AnomTrans [19] 0.9296 NCAD [22] 0.7227 TranAD [16] 0.9302 DIF [11] 0.9419 TimesNet [24] 0.8885 DCdetector [20] 0.6618 WPS [18] 0.9289 SimAD [23] 0.9417 CBMAD 0.9825 Methods Total Parameters TcnED [13] 18.71 K AnomTrans [19] 4770.85…”
InfluentialKeAD: Knowledge-enhanced Graph Attention Network for Accurate Anomaly Detection
Li, Y., Zhou, Z., Sun, P., Deng, S., Sun, X., Xue, X., Yangui, S., & Gaaloul, W. (2025) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2025.3583241 · Cited by 3
“Currently, task dependency learning for normal samples and anomaly scoring has been widely adopted for anomaly detection [8], [9], [10].”
MLAD: A Multi-Task Learning Framework for Anomaly Detection
Li, K., Tang, Z., Liang, S., Li, Z., & Liang, B. (2025) · Sensors · doi:10.3390/s25134115 · Cited by 6
“(2) Deep reconstruction models: AutoEncoder and USAD [31] are unsupervised models that learn to reconstruct normal patterns, detecting anomalies via reconstruction error.”
TriP-LLM: A Tri-Branch Patch-Wise Large Language Model Framework for Time-Series Anomaly Detection
Yu, Y.-C., Ouyang, Y., & Lin, C.-A. (2025) · IEEE Access · doi:10.1109/access.2025.3613663 · Cited by 3
“To enable the model to better capture temporal dynamics within these local patches in a single channel, the Channel Independence (CI) strategy is often applied.”
AI Anomaly Detection for Cloudified Mobile Core Architectures
Michelinakis, F., Roig, J. S. P., Malacarne, S., Xie, M., Dreibholz, T., Majumdar, S., Poe, W. Y., Patounas, G., Guerrero, C., Elmokashfi, A. M., et al. (2022) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2022.3203246 · Cited by 4
“In [32], a novel anomaly detection solution for network data is provided, where authors proposed an autoencoder architecture leveraging two decoders, and use adversarial training to identify anomalies.”
MethodLi, Q., Yan, T., Yuan, H., & Xia, Y. (2022) · Chinese Control Conference (CCC) · doi:10.23919/ccc55666.2022.9902551 · Cited by 3
Detecting Cyber Attacks in Industrial Control Systems Using Spatio-Temporal Autoencoder
Lan, B., & Yu, S. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191873 · Cited by 3
Anomaly detection and prediction evaluation for discrete nonlinear dynamical systems
Spoor, J. M., Weber, J., & Ovtcharova, J. (2023) · Transactions of the Institute of Measurement and Control · doi:10.1177/01423312231203030 · Cited by 3
“Approaches using unsupervised neural network architectures, for example, an Autoencoder as implemented by Audibert et al. (2020), require a prior training phase and the assumption of a training data set without or with only very small amount of anomalies.”
Dynamic Graph Learning with Long and Short-Term for Multivariate Time Series Anomaly Detection
Tian, Y., Gao, R., Yan, L., Liu, D., & Ye, Z. (2023) · IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS) · doi:10.1109/idaacs58523.2023.10348896 · Cited by 3
“USAD [7] applied adversarial training methods to autoencoder and proposed an autoencoder-based adversarial training network for anomaly detection.”
MethodLong, J., Luo, C., Chen, R., Yu, J., & Li, K. (2023) · The Journal of Supercomputing · doi:10.1007/s11227-023-05833-9 · Cited by 3
A weakly supervised time series anomaly detection method with dual-association discrepancy
Liu, F., Zhang, L., Li, H., Zhou, S., & Zhou, Y. (2024) · Signal Image and Video Processing · doi:10.1007/s11760-024-03119-y · Cited by 3
Bajpai, S., Krishna Murthy, P., & Kumar, N. (2024) · Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks · doi:10.1145/3643833.3656130 · Cited by 3
Chen, L., Shi, X., Zhou, L., Wang, Y., Ma, C., & Zhu, W. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-5498-4_3 · Cited by 3
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.