Time-Series Multi-Instance Learning for Weakly Supervised Industrial Fault Detection
Abstract
Time-series anomaly detection plays a crucial role in industrial fault detection. Most existing studies follow either an unsupervised setting, which is prone to false alarms, or a supervised setting, which is time-consuming and labor-intensive. To address these limitations, we adopt an innovative weakly supervised paradigm for industrial fault detection, where segment-level labels are provided during training, while point-level predictions are made during inference. Within this paradigm, we propose an innovative$C$-ary tree-based multi-instance learning (MIL) framework. First, the entire time series is represented as a$C$-ary tree, where nodes representing subsequences of different lengths are treated as instances in the MIL framework. This design allows for the detection of both point and collective anomalies. Second, to detect out-of-distribution (OOD) anomalies that are not visible during training, we develop a vector quantization module to memorize regular historical patterns. OOD anomalies are then detected when they show significant discrepancies from all memorized patterns. Finally, we enhance the MIL framework with an attention-based pooling mechanism that allocates greater focus on anomalous instances, further improving detection performance. To validate the effectiveness of our method, we conduct experiments on four real-world industrial time-series datasets. The results show that our method outperforms existing approaches by at least 6.01% in AUROC under weak supervision.
Integrity signals
Registry
Cites
Statistics
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?
Other sensor-correlation-based models make correlations explicit through graph structures, digital twins, energy-based physics features, or multi-instance representations.
“while others make correlations explicit through graph structures, digital twins, energy-based physics features, or multi-instance representations [94,96,99,101,104,106,107].”
From AI-Based Anomaly Detection in Industrial Control and Cyber–Physical Systems: A Data-Type-Oriented Systematic Review (2025) · cites this paper as [96]
Correlation-based methods are particularly suitable for attacks and faults that selectively perturb a subset of variables while keeping other measurements within nominal ranges, and for scenarios where the key symptom is a collapse or distortion of inter-sensor correlation rather than a large deviation of a single signal.
“This makes them particularly suitable for attacks and faults that selectively perturb a subset of variables while keeping other measurements within nominal ranges, as well as for scenarios where the key symptom is a collapse or distortion of inter-sensor correlation rather than a large deviation of a single signal [94,96,99-101,106-108].”
From AI-Based Anomaly Detection in Industrial Control and Cyber–Physical Systems: A Data-Type-Oriented Systematic Review (2025) · cites this paper as [96]
By focusing on joint behavior across multiple sensors and control channels, correlation-based detectors can better separate cyber-physical attacks (which often produce inconsistent or coordinated manipulation patterns) from natural faults (which follow typical failure correlations or fault signatures).
“By focusing on joint behavior across multiple sensors and control channels, correlation-based detectors can better separate cyber-physical attacks, which often produce inconsistent or coordinated manipulation patterns, from natural faults that follow typical failure correlations or fault signatures [91,95,96,100,101].”
From AI-Based Anomaly Detection in Industrial Control and Cyber–Physical Systems: A Data-Type-Oriented Systematic Review (2025) · cites this paper as [96]
The strong reliance on learned correlation profiles implies that significant operating-mode changes, reconfigurations, or sensor upgrades can cause correlation shifts that resemble attacks, requiring adaptive learning, mode-aware modeling, and careful thresholding to avoid false alarms.
“At the same time, the strong reliance on learned correlation profiles implies that significant operating-mode changes, reconfigurations, or sensor upgrades can cause correlation shifts that resemble attacks, requiring adaptive learning, mode-aware modeling, and careful thresholding to avoid false alarms [91,96,98].”
From AI-Based Anomaly Detection in Industrial Control and Cyber–Physical Systems: A Data-Type-Oriented Systematic Review (2025) · cites this paper as [96]
Ensemble and weakly supervised frameworks show that combining multiple correlation views (statistical, temporal, structural) can improve robustness under label scarcity and mixed cyber/physical disturbance conditions.
“Ensemble and weakly supervised frameworks further show that combining multiple correlation views (statistical, temporal, structural) can improve robustness under label scarcity and mixed cyber/physical disturbance conditions [95,96].”
From AI-Based Anomaly Detection in Industrial Control and Cyber–Physical Systems: A Data-Type-Oriented Systematic Review (2025) · cites this paper as [96]
The dataset distribution for sensor-correlation-based methods shows heavy evaluation on SWaT, WADI, GHL, HAI, TE, and power-system benchmarks (IEEE bus systems), with additional evidence from gas-pipeline, GRFICSv2/Factory I/O, and aNormalies-based testbeds.
“Finally, the dataset distribution indicates that sensor-correlation-based methods are heavily evaluated on SWaT, WADI, GHL, HAI, TE, and power-system benchmarks (IEEE bus systems), with additional evidence from gas-pipeline, GRFICSv2/Factory I/O, and aNormalies-based testbeds [88-96,98-102,104-110].”
From AI-Based Anomaly Detection in Industrial Control and Cyber–Physical Systems: A Data-Type-Oriented Systematic Review (2025) · cites this paper as [96]
Cites
Abdi, H., & Williams, L. J. (2010) · WIREs Computational Statistics · doi:10.1002/wics.101 · Cited by 11,104 · see citing works
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 · see citing works
LOF: identifying density-based local outliers
Breunig, M. M., Kriegel, H.-P., Ng, R. T., & Sander, J. (2000) · Proceedings of the 2000 ACM SIGMOD international conference on Management of data · doi:10.1145/342009.335388 · Cited by 4,149 · see citing works
Neural Discrete Representation Learning
Oord, A. V. D., Vinyals, O., & Kavukcuoglu, K. (2017) · arXiv · doi:10.48550/arxiv.1711.00937 · Cited by 2,729 · see citing works
Real-World Anomaly Detection in Surveillance Videos
Sultani, W., Chen, C., & Shah, M. A. (2018) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2018.00678 · Cited by 2,248 · see citing works
Ruff, L., Vandermeulen, R. A., Görnitz, N., Deecke, L., Siddiqui, S. A., Binder, A., Müller, E., & Kloft, M. (2018) · International Conference on Machine Learning · Cited by 1,719 · see citing works
Graph Neural Network-Based Anomaly Detection in Multivariate Time Series
Deng, A., & Hooi, B. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i5.16523 · Cited by 1,413 · see citing works
A Review on Outlier/Anomaly Detection in Time Series Data
Blázquez-García, A., Conde, A., Mori, U., & Lozano, J. A. (2022) · ACM Computing Surveys · doi:10.1145/3444690 · Cited by 1,102 · see citing works
SWaT: a water treatment testbed for research and training on ICS security
Mathur, A. P., & Tippenhauer, N. O. (2016) · International Workshop on Cyber-Physical Systems for Smart Water Networks (CySWater) · doi:10.1109/cyswater.2016.7469060 · Cited by 974 · see citing works
Time-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 · see citing works
Anomaly Transformer: Time Series Anomaly Detection with Association DiscrepancyPreprint
Xu, J., Wu, H., Wang, J., & Long, M. (2021) · arXiv (Cornell University) · doi:10.4230/lipics.time.2024.17 · Cited by 616 · see citing works
Ahmed, C. M., Palleti, V. R., & Mathur, A. P. (2017) · International Workshop on Cyber-Physical Systems for Smart Water Networks (CySWater) · doi:10.1145/3055366.3055375 · Cited by 574 · see citing works
DCdetector: Dual Attention Contrastive Representation Learning for Time Series Anomaly Detection
Yang, Y., Zhang, C., Zhou, T., Wen, Q., & Sun, L. (2023) · Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining · doi:10.1145/3580305.3599295 · Cited by 387 · see citing works
Interaction-Aware Graph Neural Networks for Fault Diagnosis of Complex Industrial Processes
Chen, D., Liu, R., Hu, Q., & Ding, S. X. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3132376 · Cited by 243 · see citing works
Real-World Anomaly Detection by Using Digital Twin Systems and Weakly Supervised Learning
Castellani, A., Schmitt, S., & Squartini, S. (2021) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2020.3019788 · Cited by 181 · see citing works
Improved LSTM-Based Time-Series Anomaly Detection in Rail Transit Operation Environments
Wang, Y., Du, X., Lu, Z., Duan, Q., & Wu, J. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3164087 · Cited by 161 · see citing works
Deep Weakly-supervised Anomaly Detection
Pang, G., Shen, C., Jin, H., & van den Hengel, A. (2023) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3580305.3599302 · Cited by 145 · see citing works
Filonov, P., Lavrentyev, A., & Vorontsov, A. (2016) · arXiv · doi:10.48550/arxiv.1612.06676 · Cited by 140 · see citing works
Imputation-based Time-Series Anomaly Detection with Conditional Weight-Incremental Diffusion Models
Xiao, C., Gou, Z., Tai, W., Zhang, K., & Zhou, F. (2023) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3580305.3599391 · Cited by 122 · see citing works
Graph-Augmented Normalizing Flows for Anomaly Detection of Multiple Time SeriesPreprint
Dai, E., & Chen, J. (2022) · arXiv (Cornell University) · doi:10.4230/oasics.dx.2024.11 · Cited by 68 · see citing works
Graph Structure Change-Based Anomaly Detection in Multivariate Time Series of Industrial Processes
Zhang, Z., Geng, Z., & Han, Y. M. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2023.3347000 · Cited by 64 · see citing works
Deep Contrastive One-Class Time Series Anomaly Detection
Wang, R., Liu, C., Mou, X., Gao, K., Guo, X., Pin, L., Wo, T., & Liu, X. (2023) · Society for Industrial and Applied Mathematics eBooks · doi:10.1137/1.9781611977653.ch78 · Cited by 63 · see citing works
Self-citationHuang, T., Chen, P., & Li, R. (2022) · Proceedings of the ACM Web Conference 2022 · doi:10.1145/3485447.3511984 · Cited by 61 · see citing works
Explaining Aviation Safety Incidents Using Deep Temporal Multiple Instance Learning
Janakiraman, V. M. (2018) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3219819.3219871 · Cited by 55 · see citing works
TOTEM: TOkenized Time Series EMbeddings for General Time Series AnalysisPreprint
Talukder, S. J., Yue, Y.-S., & Gkioxari, G. (2024) · Trans. Mach. Learn. Res. · doi:10.48550/arxiv.2402.16412 · Cited by 54 · see citing works
STAGED: A Spatial-Temporal Aware Graph Encoder–Decoder for Fault Diagnosis in Industrial Processes
Li, S., Meng, W., He, S., Bi, J., & Liu, G. (2023) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2023.3281083 · Cited by 45 · see citing works
Self-citation(2024) · doi:10.1109/tii.2023.3288226 · Cited by 27 · see citing works
Reward Once, Penalize Once: Rectifying Time Series Anomaly Detection
Doshi, K., Abudalou, S., & Yılmaz, Y. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9891913 · Cited by 26 · see citing works
Weakly Supervised Temporal Anomaly Segmentation with Dynamic Time Warping
Lee, D., Yu, S., Ju, H., & Yu, H. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00726 · Cited by 18 · see citing works
Liu, C., He, S., Liu, H., & Li, S. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447536 · Cited by 8 · see citing works
Self-citationNi, J., Guinet, G., Jiang, P., Callot, L., & Kan, A. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2401.10338 · Cited by 1 · see citing works
Flagged first, then by citations
1 of the 32 references couldn’t be matched to a known paper.
Cited by
Sorted by citations
Seo, J. K., Lee, J., Kim, B., Shim, W., & Seo, J. T. (2025) · Electronics · doi:10.3390/electronics15010020 · Cited by 15 · see citing works
Peng, Z., Huang, X., Fan, X., Chen, X., Feng, Z., Gan, M., Li, R., & Liao, Z. (2026) · Chemometrics and Intelligent Laboratory Systems · doi:10.1016/j.chemolab.2026.105692 · Cited by 3 · see citing works
He, Z., Yue, D., Dou, C., Yu, L., Yan, T., Li, H., & Yang, J. (2025) · IEEE Transactions on Smart Grid · doi:10.1109/tsg.2025.3636191 · Cited by 1 · see citing works
Sun, J., Fei, X., Su, L., Wang, K., Xu, Z., & Shen, H. (2026) · Structural Health Monitoring · doi:10.1177/14759217261430713 · Cited by 1 · see citing works
A Lightweight Mechanical Fault Diagnosis Network Driven by Cooperative Neuroevolution Algorithm
Yue, H., Zhang, J., Wang, Y., Tian, X., Zhang, B., Zhang, Z., & Shao, S. (2025) · Evolutionary Intelligence · doi:10.1007/s12065-025-01096-w
Wan, S., Zhou, D., Zhang, H., Zhao, M., & Chen, Y. (2026) · Journal of Building Engineering · doi:10.1016/j.jobe.2026.116592
Lin, F., Gu, S., Zhao, Z., Zhang, X., & Chen, X. (2026) · Aerospace Science and Technology · doi:10.1016/j.ast.2026.113758
Santhosh, C., Sivaraj, R., & Vijayakumar, M. (2026) · Iranian Journal of Science and Technology Transactions of Electrical Engineering · doi:10.1007/s40998-026-01258-1
- Authors
- Liu, C., He, S., Li, S., Shi, Z., & Meng, W.
- Affiliations
- State Key Laboratory of Industrial Control Technology (CN)
- Venue
- IEEE Transactions on Industrial Informatics
- Published
- 2025-04
- Volume
- 21
- Issue
- 4
- Pages
- 3326-3335
- ISSN
- 1551-3203 (ISSN-L)
- 1941-0050
- OpenAlex
- W4406856657
- Cited by
- 8
- References
- 32
- Citations by year
- 2026: 5 · 2025: 3
- FWCI
- 3.24
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Transactions on Industrial Informatics (publishedVersion)
- Funders
- National Natural Science Foundation of China (U23A20326, 62088101)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Anomaly Detection Techniques and Applications
- Physical Sciences › Engineering › Control and Systems Engineering › Fault Detection and Control Systems
- Physical Sciences › Engineering › Control and Systems Engineering › Machine Fault Diagnosis Techniques
- Keywords
- time series anomaly detection, industrial fault detection, weakly supervised learning, multiple instance learning, vector quantization, attention-based pooling
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
