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 601–700 of 1,000
MRDiff: Time Series Anomaly Detection Using Multi-level Reconstruction Diffusion
Na, D., & Kwon, J. (2024) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw65004.2024.00095 · Cited by 3
Harmful data enhanced anomaly detection for quasi-periodic multivariate time series
Wang, L., Zhou, Y. Z., Ke, W., Zheng, D., Min, F., & Li, H. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06461-x · Cited by 4
GDet: Leveraging Graph Neural Networks for Intelligent Detection of Trade-Based Market Manipulation
Xu, J., & Lin, J. (2024) · International Conference on Big Data, Artificial Intelligence and Risk Management (ICBAR) · doi:10.1145/3718751.3718866 · Cited by 2
“The training is optimized by minimizing the reconstruction error as: Finally, we select , the reconstruction error generated by decoder as the final output, which represents the part of anomalies that are difficult for the model to predict, forming a more targeted new feature for our task [14, 15].”
Anomaly Detection Method for Meteorological Data Based on Neural Networks
Guan, A., Zhu, C., Kong, L., Nan, Y., Lu, P., Liu, S., & Zhao, Y. L. (2025) · Russian Meteorology and Hydrology · doi:10.3103/s1068373924600338 · Cited by 1
Huang, X., Zhou, F., Deng, T., Chen, F., Chen, J., & Chen, N. (2025) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001425540084 · Cited by 1
A temporal dependency preserving approach for anomaly detection on multivariate time series
Benkabou, S.-E., Benabdeslem, K., Mansouri, D. E. K., Chaib, S., Mesmoudi, A., & Hadjali, A. (2025) · World Wide Web · doi:10.1007/s11280-025-01346-y · Cited by 2
Anomaly Detection for Track Dynamic Inspection Data by Information Fusion Attention and TransDetect
Zhang, Y., Yin, H., Liu, X., Huang, H., & Hou, X. (2025) · Signal Image and Video Processing · doi:10.1007/s11760-025-04305-2 · Cited by 1
“We extensively compare our model with eight baselines, including: LSTM (2018) [15], Omni-Anomaly (2019) [16], USAD (2020) [36], GDN (2020) [24], TranAD (2022) [17], AnomalyTrans (2022) [18], DCdetector (2023) [37], Wap-TAD* (2024) [38].”
Chang, H., Fan, H., Han, H., Han, M., Zhou, B., & Wang, Z. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-8170-9_7 · Cited by 1
Res2coder: A two-stage residual autoencoder for unsupervised time series anomaly detection
Wang, H., Liu, Y., Yin, H., Zheng, X., Zha, Z., Lv, M., & Guo, Z. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06684-y · Cited by 2
CBAT-ASG: Adversarial Sample Generation Method Based on CBA-Transformer
Xu, L., Yao, Z., Li, H., Diao, C., Zhou, G., Zhao, D., & Ding, H. (2025) · International Conference on Computer Supported Cooperative Work in Design (CSCWD) · doi:10.1109/cscwd64889.2025.11033503 · Cited by 1
“In order to verify the attack effect of adversarial samples, this study uses the most advanced intrusion detection models GDN [4], AEED [25], USAD [26], and DNN [27]as attack detection models.”
Huang, X., Xi, L., Fan, H., Zhangl, F., Yu, X., Liu, L., Dong, M., & Guizani, M. (2025) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2025.3583116 · Cited by 2
“…un-supervised anomaly detection has made notable progress, with methods such as SLMR (prediction-based) [19], DeepSVDD (classification-based) [20], and reconstruction-based models like AE, USAD, BeatGAN, and Ganomaly [21], [22], [23], [24], offering new perspectives for ECG quality assessment.”
DDConv: Disentangled Dual Branch Convolutional Network for Anomaly Detection in Server Machines
Yao, Y., & Yang, Z.-X. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3583424 · Cited by 1
“9) USAD [13]: It trains a pair of autoencoders adversarially to detect anomalies based on reconstruction errors.”
InfluentialData-Driven Health Status Assessment of Fire Protection IoT Devices in Converter Stations
Huang, Y., Sun, T., Cheng, Y., Zhang, J., Yang, Z., & Yang, T. (2025) · Fire · doi:10.3390/fire8070251 · Cited by 1
Zurlo, G., Cornali, R., & Ronchieri, E. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-97576-9_7 · Cited by 2
You, P., Wang, X., Chen, P., Wu, L., Chen, J., Li, X., Zeng, S., & Gao, H. (2025) · The Journal of Supercomputing · doi:10.1007/s11227-025-07665-1 · Cited by 1
“USAD [23]: uses an autoencoder with two decoders and an adversarial training framework to classify normal and abnormal data.”
Meng, D., Zhou, X., Lan, S., Wang, W., Yuan, W., & Lu, R. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-9894-3_42 · Cited by 1
Liu, B., Tao, L., Chen, X., & Li, Z. (2025) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2025.3600004 · Cited by 3
“Additionally, Generative Adversarial Network (GAN)-based approaches [27], [28], [29] employ adversarial regularization to learn temporal normal patterns for anomaly detection.”
Bridging Edge and Cloud: A Knowledge-Enhanced Framework for Efficient Time Series Anomaly Detection
Zhang, S., Zhang, J., Liu, G., Chen, S., Zhao, C., Ma, M., Chen, Y., Sun, Y., & Pei, D. (2025) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2025.3610402 · Cited by 2
Song, Y., Chen, P., Xia, Y., Liu, H., Ma, Y., & Lin, X. (2025) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws67624.2025.00105 · Cited by 1
NADHS: Online Anomaly Detection for High-Dimensional Data Streams
Ma, Q., Li, Q., Liu, X., Khoshnobish, T. I., Bai, M., Wang, X., Ning, B., & Guo, S. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3617399 · Cited by 1
“8) USAD [25] (2020s): The USAD method is based on an autoencoder architecture whose learning is inspired by GAN [47].”
Bayesian-Adaptive Graph Neural Network for Anomaly Detection (BAGNN)
Ding, Y. S., Zhang, C., Tang, S., Yang, C., & Liang, H. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-3543-9_26 · Cited by 1
Yang, W., Ou, W., Li, W., & Wang, H. (2025) · Journal of Network and Systems Management · doi:10.1007/s10922-025-09995-0 · Cited by 3
A novel time-frequency augmented perception method for multivariate time series anomaly detection
Wang, X., Ma, W., & Li, Z. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06966-5 · Cited by 1
Sanjalawe, Y. K., Fraihat, S., Al-E’mari, S., & Makhadmeh, S. N. (2025) · Annals of Data Science · doi:10.1007/s40745-025-00665-1 · Cited by 1
AttentionMamba: A Hybrid Model Combining Mamba and Transformer for Time Series Anomaly Detection
Shen, M., & Wang, G. (2025) · International Conference on Electronic Information Engineering and Computer Science (EIECS) · doi:10.1109/eiecs67708.2025.11283197 · Cited by 1
KGL: An Efficient Time Series Anomaly Detection Approach for the Industrial Internet of Things
Wang, B., Tao, Y., Wang, Y., & Zhou, L. (2025) · International Symposium on Machine Learning and Social Computing · doi:10.1145/3778450.3778526 · Cited by 1
Contrastive Learning for Robust Time Series Anomaly Detection in Cyber-Physical Systems
Ma, Y., Wang, H., Guan, H., Tao, X., & Yin, H. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-9849-3_29 · Cited by 1
Time Series Industrial Data Anomaly Detection Algorithm Based on Transformer
Li, G., Zhao, D., Li, Y., Sun, T., Lv, F., & Sun, R. (2025) · Youth Academic Annual Conference of Chinese Association of Automation (YAC) · doi:10.1109/yac66630.2025.11149870 · Cited by 1
“USAD [1] uses an adversarial training framework with dual decoders for training, allowing the model to quickly isolate anomalies while maintaining stable training.”
Liu, J., Zhou, Q., Chen, X., Zhou, J., & Li, K. (2025) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2025.3625955 · Cited by 1
Lu, J., Fang, Y., Fan, S., & Wu, X. (2025) · Journal of Intelligent & Fuzzy Systems · doi:10.1177/18758967251407728 · Cited by 2
“USAD (Audibert et al., 2020) Dual AE + adversarial Amplifies anomalies; unstable training.”
Cho, K., Lee, J., & Kim, S. B. (2026) · Information Sciences · doi:10.1016/j.ins.2026.123094 · Cited by 2
Geng, M.-Y., Li, Z., Xu, Y.-H., Liu, S.-L., Ai, Y.-B., & Zhang, W.-D. (2026) · Journal of Iron and Steel Research International · doi:10.1007/s42243-025-01692-1 · Cited by 1
Attentive LSTM-Driven Dynamic Graph Neural Networks for Multivariate Time Series Anomaly Detection
Li, Y., Zhou, Y., Yan, Q., Bo, Y., Wu, X., & Zou, W. (2025) · International Conference on Cyber-Physical Social Intelligence (ICCSI) · doi:10.1109/cpsi66656.2025.11343958 · Cited by 1
“We also evaluate several reconstruction-based methods: the hybrid LSTM-VAE [17]; DAGMM [18], which uses a Gaussian Mixture Model; the adversarially trained autoencoder USAD [19]; and the GAN-based MAD-GAN [11].”
Robust ICS Anomaly Detection Using Multi-Scale Temporal Dependencies and Frequency-Domain Features
Wang, F., Chen, H., Wang, S., Qin, Z., & Dong, F. (2026) · Electronics · doi:10.3390/electronics15030571 · Cited by 2
FST-AD: Anomaly Detection for Cyber-Physical Systems via Frequency-Spatio-Temporal GNNs
Chen, Z., Bian, X., Yang, M., Liu, C., Wu, X., & Lu, S. (2025) · IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom) · doi:10.1109/trustcom66490.2025.00270 · Cited by 1
“The five state-of-the-art frameworks for MTSAD include LSTM-VAE [23], UnSu-pervised Anomaly Detection on Multivariate Time Series (USAD) [24], Multivariate Time Series Anomaly Detection via GAT (MTSAD-GAT) [16], GDN [15], and CST-GL [4].”
An Empirical Analysis of Score Mapping in Time-Series Anomaly Detection
Park, S., Lee, G., & Ko, Y. (2025) · International Conference on Information and Communication Technology Convergence (ICTC) · doi:10.1109/ictc66702.2025.11387870 · Cited by 1
“Time-Series Anomaly Detection (TSAD) is an essential task that aims to identify abnormal data points in time-series data [3]–[7].”
EQA-MDL: Wearable ECG Signal Quality Assessment via Multi-Scale Difference Learning
Fan, H., Luo, J., Chang, H., Han, H., Tan, J., Guo, C., Muwei, L. I., & Wang, Z. (2026) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2026.3667883 · Cited by 1
Zhu, L. H., Ma, S., Xu, C., Xu, X., & Du, H. (2026) · Applied Intelligence · doi:10.1007/s10489-026-07171-8 · Cited by 1
“Audibert et al. proposed a GAN-based multivariate time-series detection algorithm, which improves model stability through staged training but lacks a targeted design for multivariate modeling [30].”
TS-TFMAE: time series self-supervised learning with time–frequency masked autoencoders
Liu, S., Li, H., & Yang, Y. (2026) · Applied Intelligence · doi:10.1007/s10489-026-07145-w · Cited by 1
“USAD [26 ] builds an autoencoder model with two encoders and one decoder and then adds adversarial training to improve the representation capacity.”
An Efficient Real-Time Anomaly Detection Scheme for Water Quality Monitoring
Guo, W., Huang, L., Li, Y., & Luo, W. (2026) · Water · doi:10.3390/w18060726 · Cited by 1
Pseudo Anomalies and Hard Sample Mining for Ventricular Arrhythmia Anomaly Detection
Fan, H., Gao, S., Han, H., Han, M., Zhou, B., & Wang, Z. (2026) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2026.3676678 · Cited by 2
Industrial chain risk monitoring with explainable dynamic knowledge graph learning
Guo, Y., Ma, Y., Deng, S., Lv, Z., & Wang, X. (2026) · Array · doi:10.1016/j.array.2026.100796 · Cited by 1
RFETR: Residual Frequency-Enhanced Two-stage Reconstruction for Time Series Anomaly Detection
Jing, S., Wu, J., Zhan, P., & Zhang, T. (2026) · International Conference on Communication Networks and Machine Learning (CNML) · doi:10.1109/cnml68938.2026.11453190 · Cited by 1
TransEdge: a lightweight transformer-based anomaly detection framework for edge devices
Shi, Q., Yang, Z., & Shao, S. (2026) · Journal of Intelligent Information Systems · doi:10.1007/s10844-026-01043-w · Cited by 4
“…into window-level decisions via post-processing, a conventional per-timestamp confusion-matrix diagram is less infor - mative for the adjusted scores; therefore, we report precision, recall, and F 1 -score to ensure fair comparability with prior work (Audibert et al., 2020; Qiu et al., 2025).”
InfluentialLiu, H., Huang, X., Jia, M., Jia, T., Wu, Z., Li, Y., Wang, Y., Tan, Y., Zhu, L., & Zhou, W. (2026) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2026.3684516 · Cited by 1
“Reconstruction-based methods [1], [17], [18], [20], [21], [32], [33], [34], [35] posit that by learning to reconstruct unlabeled training data, they can fit the distribution of normal data, accurately reconstructing normal samples while struggling to reconstruct anomalous ones.”
Yu, Z., Li, J. W., Sun, L., & Jiang, Q. (2026) · Transactions of the Institute of Measurement and Control · doi:10.1177/01423312261439445 · Cited by 1
CUTIE: Component-specific Unsupervised Technique for In-node Examination
Vignali, A., & d’Ambrosio, N. (2026) · Computer Networks · doi:10.1016/j.comnet.2026.112339 · Cited by 1
Aslam, M. M., Tufail, A., Ding, Y., Silva, L. C. D., Apong, R. A. A. H. M., & Zuhairi, M. F. (2026) · Technologies · doi:10.3390/technologies14050267 · Cited by 1
Luo, Z., Sheng, J., Xu, L., & Yi, Y. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.113924 · Cited by 1
Dabthong, H., Ketcham, M., & Boonyopakorn, P. (2026) · IEEE International Conference on Cybernetics and Innovations (ICCI) · doi:10.1109/icci68752.2026.11506458 · Cited by 1
“USAD [6] employs dual autoencoders on SWaT and WADI, achieving F1 = 0.846 on SWaT.”
InfluentialADAD: Multivariate Time Series Anomaly Detection via Adaptive Series Decomposition
Yao, Y., & Yang, Z.-X. (2026) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2026.3695421 · Cited by 1
MSMGNet: Multiscale Modeling of Interseries Dependency for Time Series Anomaly Detection
Wang, S., Zhou, Y., Wei, Y., & Yong, P. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3697065 · Cited by 1
“To ensure a fair comparison, we adopt the widely used point-adjustment strategy [38].”
Zhong, B., Li, J., Pu, Z., & Zhang, R. (2026) · Applied Sciences · doi:10.3390/app16115450 · Cited by 1
AnomalyFilter: Selective Denoising Diffusion Model for Time Series Anomaly Detection
Obata, K., Chen, Z., Matsubara, Y., Zhu, L., & Sakurai, Y. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-1300-9_34 · Cited by 1
Alhadidi, M., & Alkasassbeh, M. S. (2026) · International Conference on Computational Intelligence Approaches and Applications (ICCIAA) · doi:10.1109/icciaa68481.2026.11543640 · Cited by 1
Lee, M., Lin, J., & Katsikas, S. K. (2026) · ACM Symposium on Applied Computing · doi:10.1145/3748522.3779783 · Cited by 1
“Audibert et al. [2] proposed USAD, an unsupervised anomaly detection method for multivariate time series that leverages au-toencoders with adversarial training.”
Hartung, F., Muraleedharan, A., Kloft, M., & Burger, J. (2026) · Systems and Control Transactions · doi:10.69997/sct.127956 · Cited by 1
Alatawi, M. N. (2026) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-026-01252-9 · Cited by 2
“In recent years, several methods have emerged that exploit advances in more sophisticated models to better capture the temporal dependencies of sequential data: TranAD by Tuli et al. [17] and USAD by Audibert et al. Int J Comput Intell Syst https://doi.org/10.1007/s44196-026-01252-9 [ 18].”
InfluentialCommunity self-organization based multi-agent collaboration for industrial sensor fault detection
Cheng, J., Pei, S., Luo, B., & Zhang, S. (2026) · Measurement Science and Technology · doi:10.1088/1361-6501/ae81d7 · Cited by 1
“The proposed C 2 MAD is compared with several representative anomaly detection methods, including USAD [24], OA [25], LSTM-DT [26], AT [27], TimesNet [3], and FITS [28].”
Yu, R., Wang, M., Yun, J., Du, J., Wang, L., & Suo, Y. (2026) · International Conference on Computing, Networks and Internet of Things (CNIOT) · doi:10.1109/cniot69547.2026.11582255 · Cited by 1
VQFlow: A Benchmark Dataset for Encrypted Video Streaming Traffic across QoS Configurations
Zhao, Z., Li, Z., Li, T., & Zhang, F. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770854.3785690 · Cited by 1
“• USAD (Unsupervised Anomaly Detection) [54]: a state-of-the-artunsupervisedframeworkthat combinesautoencoder-based reconstruction with adversarial training, enabling robust detection of anomalies in time series data.”
Global-key region collaborative noise-aware for electrocardiogram signal quality assessment
Fan, J., Chang, H., Han, H., & Luo, J. (2026) · PLoS ONE · doi:10.1371/journal.pone.0355335 · Cited by 2
“• USAD [33]: An adversarial autoencoder with one encoder and two decoders, providing stable training.”
Nouman, M., Saleh, M. H., & Ujjan, R. M. A. (2026) · Future Internet · doi:10.3390/fi18090447 · Cited by 1
IoT-GAN: Anomaly Detection for Time Series in IoT Based on Generative Adversarial Networks
Chen, X., Zhang, S., Jiang, Q., Chen, J., Huang, H., & Gu, C. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-95388-1_45 · Cited by 1
Chen, B., Zhao, X., 代, 良., 陈, 渤., 刘, 畅., & 赵, 小. (2022) · Scientia Sinica Informationis · doi:10.1360/ssi-2022-0277 · Cited by 1
MTAP-DK: Multivariate Time-Series Anomaly Prediction with Domain Knowledge
Xue, L., Peng, Z., Zhang, J., Wang, F., & Wang, Y. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892923 · Cited by 2
“Some methods [1], [13] reconstruct the multivariate time series and diagnose anomalies based on the reconstruction error.”
BackgroundMethodAn Unsupervised Convolutional Adversarial Anomaly Detection Model for IoT Data Infrastructure
Wen, P., Yang, Z., Wu, L., Qi, S., & Chen, J. (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-dependsys57074.2022.00208 · Cited by 1
“Moreover, the static threshold can not effectively detect “slight” anomalies when detecting anomalies in the collected data, so “slight” anomalies are also a current research focus [2].”
BackgroundMethodInfluentialSTAD: Multivariate Time Series Anomaly Detection Based on Spatio-Temporal Relationship
Chen, K., Zhao, G., Yao, Z., & Zhang, Z. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-46661-8_6 · Cited by 1
Miyake, R., Tominaga, S., Terakado, Y., Takado, N., Aoki, T., Miyamoto, C., Naito, S., Taguchi, Y., Kato, Y., & Nakata, K. (2023) · Journal of Nuclear Engineering and Radiation Science · doi:10.1115/1.4064123 · Cited by 1
Empirical Analysis of the Fine-Tuning for Unsupervised Anomaly Detection in the ICT System
Matsuo, Y. (2023) · International Conference on Network and Service Management (CNSM) · doi:10.23919/cnsm59352.2023.10327845 · Cited by 1
Li, Z., Xu, D., Li, Y., Chai, T., & Yang, T. (2023) · IEEE Symposium Series on Computational Intelligence (SSCI) · doi:10.1109/ssci52147.2023.10371977 · Cited by 1
“However, the data in application fields is large, and the proportion of abnormal data is small, therefore the cost of labeling all data is prohibitively high [8], [9].”
BackgroundMulti-view Deep Markov Models for Time Series Anomaly Detection
Nguyen, P., Tran, H., & Le, T. M. V. (2023) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata59044.2023.10386155 · Cited by 1
Ma, H., Wu, Y., Liu, F., Zhang, S., Shi, Y., Ma, S., & Liu, M. (2023) · International Conference on Image Processing, Computer Vision and Machine Learning (ICICML) · doi:10.1109/icicml60161.2023.10424770 · Cited by 1
“The experiment selects the anomaly detection method of the industrial Internet of Things time series that has achieved excellent performance in the past three years USAD [7] , MTAD-GAT [8] , HNN-NTL [9] , GTA [10] , LSTM-Attention [11] , RTransformer [12] .”
MethodDGFormer: An Effective Dynamic Graph Transformer Based Anomaly Detection Model for IoT Time Series
He, H., Li, X., Chen, P., Chen, J., Song, W., & Xi, Q. (2024) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-031-54528-3_10 · Cited by 3
Transformer-Based Method for Unsupervised Anomaly Detection of Flight Data
Yu, H., Wu, H., Sun, Y., & Liu, H. (2024) · Lecture notes in electrical engineering · doi:10.1007/978-981-97-3998-1_142 · Cited by 1
PaSTS An Operational Dataset for Domestic Solar Thermal Systems
Ebmeier, F., Ludwig, N., Martius, G., & Franz, V. H. (2024) · ACM International Conference on Future Energy Systems (e-Energy) · doi:10.1145/3632775.3662159 · Cited by 1
Point-Correlate Adversarial Transformer for Unsupervised Multivariate Time Series Anomaly Detection
Li, H., Kong, X., Shen, G., Yan, X., Yang, Y., & Collotta, M. (2024) · International Conference on Computer Supported Cooperative Work in Design (CSCWD) · doi:10.1109/cscwd61410.2024.10580188 · Cited by 1
“Audibert et al. [9] proposed an anomaly detection structure based on adversely trained autoencoders called UnSupervised Anomaly Detection for multivariate time series (USAD).”
MethodAnomaly Detection Method for Multivariate Time Series Data Based on BLTranAD
Zhang, C., Wu, S., Gao, M., Li, Y., Shi, G., Li, Y., & Ma, H. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-5618-6_2 · Cited by 1
Zhou, K., Wei, H., & Wang, W. (2024) · International Conference on Electrical Engineering, Control and Robotics (EECR) · doi:10.1109/eecr60807.2024.10607164 · Cited by 1
Improving Air Quality Data Analysis by Injecting and Detecting Contextual Anomalies
Emvoliadis, A., Angelis, G.-F., Drosou, A., & Tzovaras, D. K. (2024) · IEEE International Geoscience and Remote Sensing Symposium (IGARSS) · doi:10.1109/igarss53475.2024.10642679 · Cited by 1
Industrial Anomaly Detection Based on Factorized Temporal-Channel Fusion
Xing, Y., Liu, D., & Bao, Y. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3452955 · Cited by 1
“Using windows to replace the original data as input is quite common in previous works [25], [33], [34].”
BackgroundMethodChen, C., & Chang, S. (2024) · International Conference on Control, Decision and Information Technologies (CoDIT) · doi:10.1109/codit62066.2024.10708222 · Cited by 1
Ranking Neighborhood and Class Prototype Contrastive Learning for Time Series
Wei, C., Yuan, J., Zhang, Y., Yu, Z., Liu, Y., & Liu, H. (2024) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2024.3495509 · Cited by 1
“To demonstrate the performance of RESEAL, we compare it with TS2Vec, DAGMM [36], AnomalyBERT [37], USAD [38], GDN[39],andTranAD[40]forthedetectionofMTSanomalies.”
Zhang, Y., & Gong, C. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0811-9_18 · Cited by 1
Nahalka, M., Cook, M. M., & Pezaros, D. P. (2024) · IFIP advances in information and communication technology · doi:10.1007/978-3-031-81900-1_2 · Cited by 3
In-Vehicle Network Anomaly Detection Based on a Graph Attention Network
Luo, F., Luo, C., Wang, J., & Li, Z. (2025) · SAE International Journal of Connected and Automated Vehicles · doi:10.4271/12-08-04-0034 · Cited by 1
Multivariable time series anomaly detection based on cross spatial-temporal attention mechanism
Zhou, P., Yangyang, Z. E. N. G., ZHANG, Y., JIANG, W., & Zhang, R. (2025) · Scientia Sinica Informationis · doi:10.1360/ssi-2024-0274 · Cited by 1
Graph Structure Learning via Transfer Entropy for Multivariate Time Series Anomaly Detection
Liu, M.-Y., Wang, Y., Zhou, X., & Wang, Y. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889333 · Cited by 3
“We compare the combined methods with a range of state-of-the-arts in multivariate time series anomaly detection, including Omni-Anomaly [13], MSCRED [16], TimesNet [17], USAD [18], TranAD [19], DAGMM [20], FCSTGNN [21] and Fouri-erGNN [22].”
Multivariate Time Series Data Mining for Failure Prediction & Root Cause Analysis
Agarwal, N., Gupta, G. R., Jatala, V., & Saha, A. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888297 · Cited by 1
“Most of the previous studies on anomaly detection [1] [2] [3] [4] have analyzed synthetically generated anomalies.”
Harnessing Contrastive Learning and Neural Transformation for Time Series Anomaly Detection
Chen, J., Feng, M., & Wirjanto, T. S. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890321 · Cited by 9
“2) Baseline Methods and Evaluation Metrics: We compare our approach to several recent anomaly detection meth-ods, including MAD-GAN [3], USAD [9], MSCRED [10], TranAD [11], NeuTralAD [26], AnomalyTransformer [12], COCA [30], DCdetector [31] and CST-GL [32].”
Yan, X., Yang, S., Guo, Y., Fu, Y., Wu, S., & Li, J. (2025) · Mechanisms and machine science · doi:10.1007/978-3-031-82907-9_6 · Cited by 1
Li, G., Yang, W., Tian, X., & Peng, J. (2024) · IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom) · doi:10.1109/trustcom63139.2024.00239 · Cited by 2
“Comparison experiments include LSTM-NDT [21], GDN [22], DAGMM [23], OmniAnomaly [24], MSCRED [25], MAD-GAN [26], USAD [27], MTAD-GAT [28], CAE-M [29], TranAD [30], and DTTAD [31].”
FEFM: Feature Extraction and Fusion Module for Enhanced Time Series Anomaly Detection
Jeon, S. H., Kim, K., & Choi, Y. S. (2025) · ACM Symposium on Applied Computing · doi:10.1145/3672608.3707794 · Cited by 1
Spatio-Temporal Weighted Graph Reason Learning for Multivariate Time-Series Anomaly Detection
Zhang, H., Pang, H., Zhao, Y., Gao, G., Song, Y., Wei, S., & Zhao, Y. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3569316 · Cited by 3
“Deep learning approaches, such as AutoEncoder (AE) [21], RNNs [22], CNNs [23], [24], and Transformer are commonly employed, using prediction or reconstruction errors as anomaly indicators.”
Lu, R., Zheng, D., Yang, Q., & Cao, W. (2024) · International Conference on Neuromorphic Computing (ICNC) · doi:10.1109/icnc64304.2024.10987651 · Cited by 1
Lee, S., & Kim, T. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3573336 · Cited by 5
“Recently, reconstruction methods have been extended to include adversarial training strategies [18], [19] or combined with machine learning-based techniques [20].”
Fu, N., Cheng, G., Dai, G., Mei, H., Qiu, X., & Teng, Y. (2025) · Journal of Systems and Software · doi:10.1016/j.jss.2025.112513 · Cited by 4
An Unsupervised Learning Model for Intelligent Machine‐Failure Prediction With Heterogeneous Sensors
Park, J.-H., Kim, J., Lee, D.-W., Kim, H., & Hong, D. (2025) · International Journal of Intelligent Systems · doi:10.1155/int/3346341 · Cited by 2
“Te threshold is determined by considering the maximum reconstruction error observed in a validation set composed of normal data [59, 61].”
Anomaly Detection in Vehicular Networks Using Causality-Aware Graph Convolutional Networks (CA-GCN)
Luo, F., Luo, C., Wang, J., Li, Z., Liao, Z., & Liu, Q. (2025) · International Journal of Automotive Technology · doi:10.1007/s12239-025-00305-w · Cited by 1
Jang, S., Zhao, T., Yogarathnam, A., Yue, M., Ding, F., & Zhao, J. (2025) · IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm) · doi:10.1109/smartgridcomm65349.2025.11204620 · Cited by 1
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.