A Semi-Supervised VAE Based Active Anomaly Detection Framework in Multivariate Time Series for Online Systems
The DOI matches its Crossref record; 34 of the 35 references it cites were found in a registry or an index, and 1 are not checked yet.
Abstract
Nowadays, the large online systems are constructed on the basis of microservice architecture. A failure in this architecture may cause a series of failures due to the fault propagation. Thus, the large online systems need to be monitored comprehensively to ensure the service quality. Even though many anomaly detection techniques have been proposed, few of them can be directly applied to a given microservice or cloud server in industrial environment. To settle these challenges, this paper presents SLA-VAE, a semi-supervised learning based active anomaly detection framework using variational auto-encoder. SLA-VAE first defines anomalies based on feature extraction module, introduces semi-supervised VAE to identify anomalies in multivariate time series, and employs active learning to update the online model via a small number of uncertain samples. We conduct experiments on the cloud server data from two different types of game business in Tencent. The results show that SLA-VAE significantly outperforms other state-of-the-art methods and is suitable for wide deployment in large online business system.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
In the citing paper
“Semi-supervised learning approaches aim to train models that generalize across servers and container instances without requiring per-instance training [105]”
From Anomaly detection and root-cause identification in microservices: a survey (2026) · cites this paper as [105]
Citations
Cites
Li, Z., & Hoiem, D. (2018) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2017.2773081 · Cited by 4,202
Saliency Detection: A Spectral Residual Approach
Hou, X., & Zhang, L. (2007) · 2007 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2007.383267 · Cited by 3,422 · cited 2×
“In this paper, we adopt the Spectral Residual (SR) algorithm[12] to measure residual errors.”
MethodMicroservices: Yesterday, Today, and Tomorrow
Dragoni, N., Giallorenzo, S., Lafuente, A. L., Mazzara, M., Montesi, F., Mustafin, R., & Safina, L. (2017) · Springer eBooks · doi:10.1007/978-3-319-67425-4_12 · Cited by 2,058
DeepLog: Anomaly Detection and Diagnosis from System Logs through Deep Learning
Du, M., Li, F., Zheng, G., & Srikumar, V. (2017) · Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security · doi:10.1145/3133956.3134015 · Cited by 1,835
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 5×
“When detecting anomalies in multivariate time series, the existing approaches[7, 13, 17, 20, 21, 25, 30, 33, 37, 39, 40] directly adopt a sequence of observations x t − w : t to calculate the anomaly result of x t at time t .”
BackgroundMethodInfluentialDetecting 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 · cited 5×
“When detecting anomalies in multivariate time series, the existing approaches[7, 13, 17, 20, 21, 25, 30, 33, 37, 39, 40] directly adopt a sequence of observations x t − w : t to calculate the anomaly result of x t at time t .”
BackgroundMethodInfluentialActive Learning with Statistical Models
Cohn, D. A., Ghahramani, Z., & Jordan, M. I. (1996) · Journal of Artificial Intelligence Research · doi:10.1613/jair.295 · Cited by 1,317 · cited 1×
“Second, to optimize the online detection model continuously, we adopt the active learning, which is aimed to learn and update models from small amount of data[5], and propose an active anomaly detection”
MethodTime Series FeatuRe Extraction on basis of Scalable Hypothesis tests (tsfresh – A Python package)
Christ, M., Braun, N., Neuffer, J., & Kempa-Liehr, A. W. (2018) · Neurocomputing · doi:10.1016/j.neucom.2018.03.067 · Cited by 1,257 · cited 1×
“Traditional time series feature extraction methods obtain features from statistical, temporal and spectral domains[3, 4, 26, 28].”
MethodMAD-GAN: Multivariate Anomaly Detection for Time Series Data with Generative Adversarial Networks
Li, D., Chen, D., Jin, B., Shi, L., Goh, J., & Ng, S.-K. (2019) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-30490-4_56 · Cited by 1,185
A 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
USAD: UnSupervised Anomaly Detection on Multivariate Time Series
Audibert, J., Michiardi, P., Guyard, F., Marti, S., & Zuluaga, M. A. (2020) · Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3394486.3403392 · Cited by 1,094
Zhang, C., Song, D., Chen, Y., Feng, X., Lumezanu, C., Cheng, W., Ni, J., Zong, B., Chen, H., & Chawla, N. V. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33011409 · Cited by 855
LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection
Malhotra, P., Ramakrishnan, A., Anand, G., Vig, L., Agarwal, P., & Shroff, G. (2016) · arXiv · doi:10.48550/arxiv.1607.00148 · Cited by 819 · cited 4×
“Unsupervised works use Bayesian network[7], graph neural network (GNN)[33, 39], long short-term memory (LSTM)[13, 21], LSTM-based auto-encoder[20, 25], LSTM-based generative adversarial network (GAN)[17] to forecast or reconstruct multivariate time series, and adopt prediction error or…”
BackgroundMethodInfluentialUnsupervised Anomaly Detection via Variational Auto-Encoder for Seasonal KPIs in Web Applications
Xu, H., Feng, Y., Chen, J., Wang, Z., Qiao, H., Chen, W., Zhao, N., Li, Z., Bu, J., Li, Z., et al. (2018) · Proceedings of the 2018 World Wide Web Conference on World Wide Web - WWW '18 · doi:10.1145/3178876.3185996 · Cited by 808 · cited 2×
“The existing approaches mainly detect anomalies in online systems from four perspectives, namely KPI-level (key performance indicator)[19, 36], entity-level[30, 39], log-level[9, 22], and trace-level[10, 24].”
BackgroundMethodMultivariate Time-Series Anomaly Detection via Graph Attention Network
Zhao, H., Wang, Y., Duan, J., Huang, C., Cao, D., Tong, Y., Xu, B., Bai, J., Tong, J., & Zhang, Q. (2020) · 2020 IEEE International Conference on Data Mining (ICDM) · doi:10.1109/icdm50108.2020.00093 · Cited by 743 · cited 5×
“When detecting anomalies in multivariate time series, the existing approaches[7, 13, 17, 20, 21, 25, 30, 33, 37, 39, 40] directly adopt a sequence of observations x t − w : t to calculate the anomaly result of x t at time t .”
BackgroundMethodInfluentialMicroservices Architecture Enables DevOps: Migration to a Cloud-Native Architecture
Balalaie, A., Heydarnoori, A., & Jamshidi, P. (2016) · IEEE Software · doi:10.1109/ms.2016.64 · Cited by 735 · cited 1×
“This kind of architecture have shown great advantages in independent deployment, rapid delivery, and flexible expansion[2, 8, 32].”
BackgroundLogAnomaly: Unsupervised Detection of Sequential and Quantitative Anomalies in Unstructured Logs
Meng, W., Liu, Y., Zhu, Y., Zhang, S., Pei, D., Liu, Y., Chen, Y., Zhang, R., Tao, S., Sun, P., et al. (2019) · Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2019/658 · Cited by 708 · cited 1×
“The existing approaches mainly detect anomalies in online systems from four perspectives, namely KPI-level (key performance indicator)[19, 36], entity-level[30, 39], log-level[9, 22], and trace-level[10, 24].”
BackgroundTime-Series Anomaly Detection Service at Microsoft
Ren, H., Xu, B., Wang, Y., Yi, C., Huang, C., Kou, X., Xing, T., Yang, M., Tong, J., & Zhang, Q. (2019) · Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3292500.3330680 · Cited by 623 · cited 5×
“[28] uses convolutional neural network to detect anomalies in univariate time series.”
BackgroundMethodInfluentialAnomaly Detection in Streams with Extreme Value Theory
Siffer, A., Fouque, P.-A., Termier, A., & Largouet, C. (2017) · Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/3097983.3098144 · Cited by 553 · cited 2×
“Second, we apply the peaks-over-threshold (POT)[29] to automatically select the initial threshold.”
BackgroundMethodTSFEL: Time Series Feature Extraction Library
Barandas, M., Folgado, D., Fernandes, L., Santos, S., Abreu, M., Bota, P., Liu, H., Schultz, T., & Gamboa, H. (2020) · SoftwareX · doi:10.1016/j.softx.2020.100456 · Cited by 525
Mitchell, T., Cohen, W., Hruschka, E., Talukdar, P., Yang, B., Betteridge, J., Carlson, A., Dalvi, B., Gardner, M., Kisiel, B., et al. (2018) · Communications of the ACM · doi:10.1145/3191513 · Cited by 507 · cited 1×
“In the future, we may introduce the transfer learning[18] and never ending learning[23] to learn multiple uncertain tasks and integrate knowledge over time.”
BackgroundGeneric and Scalable Framework for Automated Time-series Anomaly Detection
Laptev, N., Amizadeh, S., & Flint, I. (2015) · Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2783258.2788611 · Cited by 425 · cited 1×
“A great amount of multivariate time series anomaly detection techniques[1, 6, 7, 11, 13, 16, 17, 19–21, 25, 33, 35, 37–40] have been proposed in recent years.”
BackgroundWeb traffic anomaly detection using C-LSTM neural networks
Kim, T.-Y., & Cho, S.-B. (2018) · Expert Systems with Applications · doi:10.1016/j.eswa.2018.04.004 · Cited by 340 · cited 2×
“C-LSTM[14] extracts features by combining CNN and LSTM.”
MethodSeer: Leveraging Big Data to Navigate the Complexity of Performance Debugging in Cloud Microservices
Gan, Y., Zhang, Y., Hu, K., Cheng, D., He, Y., Pancholi, M., & Delimitrou, C. (2019) · Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems · doi:10.1145/3297858.3304004 · Cited by 269 · cited 1×
“The existing approaches mainly detect anomalies in online systems from four perspectives, namely KPI-level (key performance indicator)[19, 36], entity-level[30, 39], log-level[9, 22], and trace-level[10, 24].”
BackgroundOpprentice: Towards Practical and Automatic Anomaly Detection Through Machine Learning
Liu, D., Zhao, Y., Xu, H., Sun, Y., Pei, D., Luo, J., Jing, X., & Feng, M. (2015) · Proceedings of the 2015 Internet Measurement Conference · doi:10.1145/2815675.2815679 · Cited by 252 · cited 6×
“Based on well extracted features, [19] adopts random forest classifier for anomaly detection.”
BackgroundMethodInfluentialActive learning: theory and applications
Koller, D., & Tong, S. (2001) · Cited by 210
CloudRanger: Root Cause Identification for Cloud Native Systems
Wang, P., Xu, J., Ma, M., Lin, W., Pan, D., Wang, Y., & Chen, P. (2018) · 2018 18th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID) · doi:10.1109/ccgrid.2018.00076 · Cited by 181 · cited 1×
“In active learning, an acquisition function, that decides which observations is need to ask for a label, relies on model uncertainty[31].”
BackgroundSelf-citationDecoupling Representation Learning and Classification for GNN-based Anomaly Detection
Wang, Y., Zhang, J., Guo, S., Yin, H., Li, C., & Chen, H. (2021) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3404835.3462944 · Cited by 176
A Spatiotemporal Deep Learning Approach for Unsupervised Anomaly Detection in Cloud Systems
He, Z., Chen, P., Li, X., Wang, Y., Yu, G., Chen, C., Li, X., & Zheng, Z. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.3027736 · Cited by 148 · cited 2×
“A great amount of multivariate time series anomaly detection techniques[1, 6, 7, 11, 13, 16, 17, 19–21, 25, 33, 35, 37–40] have been proposed in recent years.”
BackgroundSelf-citationAnomaly Detection from System Tracing Data Using Multimodal Deep Learning
Nedelkoski, S., Cardoso, J., & Kao, O. (2019) · 2019 IEEE 12th International Conference on Cloud Computing (CLOUD) · doi:10.1109/cloud.2019.00038 · Cited by 139 · cited 1×
“The existing approaches mainly detect anomalies in online systems from four perspectives, namely KPI-level (key performance indicator)[19, 36], entity-level[30, 39], log-level[9, 22], and trace-level[10, 24].”
BackgroundDual channel LSTM based multi-feature extraction in gait for diagnosis of Neurodegenerative diseases
Zhao, A., Qi, L., Dong, J., & Yu, H. L. (2018) · Knowledge-Based Systems · doi:10.1016/j.knosys.2018.01.004 · Cited by 98 · cited 2×
“The authors in [38] proposed a dual channel LSTM to extract features from time data and force data.”
BackgroundMultivariate-Time-Series-Driven Real-time Anomaly Detection Based on Bayesian Network
Ding, N., Gao, H., Bu, H., Ma, H., & Si, H. (2018) · Sensors · doi:10.3390/s18103367 · Cited by 65 · cited 4×
“Unsupervised works use Bayesian network[7], graph neural network (GNN)[33, 39], long short-term memory (LSTM)[13, 21], LSTM-based auto-encoder[20, 25], LSTM-based generative adversarial network (GAN)[17] to forecast or reconstruct multivariate time series, and adopt prediction error or…”
BackgroundMethodInfluentialSDFVAE: Static and Dynamic Factorized VAE for Anomaly Detection of Multivariate CDN KPIs
Dai, L., Lin, T., Liu, C., Jiang, B., Liu, Y., Xu, Z., & Zhang, Z.-L. (2021) · Proceedings of the Web Conference 2021 · doi:10.1145/3442381.3450013 · Cited by 39 · cited 3×
“We compare SLA-VAE with four baseline models, namely USAD[1], SDFVAE[6], MSCRED[37], and OmniAD[30].”
BackgroundMethodPirasteh, P., Nowaczyk, S., Pashami, S., Löwenadler, M., Thunberg, K., Ydreskog, H., & Berck, P. (2019) · doi:10.1145/3304079.3310288 · Cited by 27
Wu, J., Yao, L., Liu, B., Ding, Z., & Zhang, L. (2021) · Advances in Mechanical Engineering · doi:10.1177/16878140211013138 · Cited by 13 · cited 6×
“MTLED[35] uses a CNN module for processing spatial information and an LSTM module for modeling temporal information without losing the time dimension, and concatenates them along the time dimension.”
BackgroundMethodInfluential
Research works printed but not linked: 6
- 10.5555/1622737.1622744 · doi:10.5555/1622737.1622744
- Diederik P Kingma and Max Welling. 2013. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114(2013). · arXiv
- Pankaj Malhotra, Lovekesh Vig, Gautam Shroff, and Puneet Agarwal. 2015. Long short term memory networks for anomaly detection in time series. In Proceedings, Vol. 89. 89–94.
- Xuming Ran Mingkun Xu Lingrui Mei Qi Xu and Quanying Liu. 2020. Detecting out-of-distribution samples via variational auto-encoder with reliable uncertainty estimation. arXiv preprint arXiv:2007.08128(2020). · arXiv
- Svante Wold, Kim Esbensen, and Paul Geladi. 1987. Principal component analysis. Chemometrics and intelligent laboratory systems 2, 1-3 (1987), 37–52.
- 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 International conference on learning representations.
Cited by
Deep Learning for Time Series Anomaly Detection: A Survey
Zamanzadeh Darban, Z., Webb, G. I., Pan, S., Aggarwal, C., & Salehi, M. (2025) · ACM Computing Surveys · doi:10.1145/3691338 · Cited by 613
“In order to prevent big losses from total machine failure or even natural disasters, it is necessary to detect OOS anomalies early enough.”
Xin, R., Chen, P., & Zhao, Z. (2023) · Journal of Systems and Software · doi:10.1016/j.jss.2023.111724 · Cited by 99
Self-citationRevisiting VAE for Unsupervised Time Series Anomaly Detection: A Frequency Perspective
Wang, Z., Pei, C., Ma, M., Wang, X., Li, Z., Pei, D., Rajmohan, S., Zhang, D., Lin, Q., Zhang, H., et al. (2024) · International World Wide Web Conference (WWW) · doi:10.1145/3589334.3645710 · Cited by 173
Wang, F., Jiang, Y., Zhang, R., Wei, A., Xie, J., & Pang, X. (2025) · Sensors · doi:10.3390/s25010190 · Cited by 119
“SLA-VAE [44] addresses the challenge of obtaining labeled data in InterFusion through semi-supervised learning, utilizing active learning to leverage a small number of highly uncertain samples for learning and updating online system anomaly detection models.”
Zheng, M., Man, J., Wang, D., Chen, Y., Li, Q., & Liu, Y. (2023) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2023.109235 · Cited by 91
Metadata
- Authors
- Huang, T., Chen, P., & Li, R.
- Affiliations
- Tencent (China) (CN)
- Sun Yat-sen University (CN)
- Venue
- Proceedings of the ACM Web Conference 2022
- Publisher
- ACM
- Published
- 2022-04-25
- Pages
- 1797-1806
- OpenAlex
- W4224315779
- Cited by
- 61
- References
- 34
- Citations by year
- 2026: 9 · 2025: 19 · 2024: 16 · 2023: 14 · 2022: 3
- FWCI
- 12.60
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- International World Wide Web Conference (WWW) (publishedVersion)
- Funders
- Key-Area Research and Development Program of Guangdong Province (No. 2020B010165002)
- Basic and Applied Basic Research of Guangzhou (No. 202002030328)
- National Key Research and Development Program of China (2019YFB1804002)
- National Natural Science Foundation of China (No. 61802448)
- Natural Science Foundation of Guangdong Province (No. 2019A1515012229)
- Tencent Rhino-Bird Joint Research Program (No.K21-67000-059)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Software System Performance and Reliability
- Physical Sciences › Computer Science › Artificial Intelligence › Anomaly Detection Techniques and Applications
- Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
- Keywords
- anomaly detection, semi-supervised learning, variational autoencoder, active learning (machine learning), multivariate time series, microservices, fault propagation
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.