Federated Learning for Multivariate Time Series Anomaly Detection in Industrial Automation

Khayyam Nosrati; Martin Uray; Olaf Sassnick; Saverio Messineo; Stefan Huber

arxiv: 2605.27486 · v1 · pith:CDFD5Z6Wnew · submitted 2026-05-26 · 💻 cs.LG

Federated Learning for Multivariate Time Series Anomaly Detection in Industrial Automation

Khayyam Nosrati , Martin Uray , Saverio Messineo , Olaf Sassnick , Stefan Huber This is my paper

Pith reviewed 2026-06-29 18:42 UTC · model grok-4.3

classification 💻 cs.LG

keywords federated learningmultivariate time seriesanomaly detectionindustrial automationcyclic dynamicsdatasetbenchmarking

0 comments

The pith

A new dataset with cyclic dynamics from repetitive industrial processes enables benchmarking of federated multivariate time series anomaly detection methods.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper identifies that current datasets for federated learning in multivariate time series anomaly detection lack the scale, accurate labels, and freedom from flaws needed for reliable evaluation, while also leaving cyclic process behavior underexplored. It addresses this by creating a dataset built around the repetitive cycles typical of discrete automation and tests selected anomaly detection methods on both this new data and an existing public benchmark. A reader would care because federated setups keep sensitive industrial data private while still allowing anomaly detection across distributed sites. If the claim holds, evaluations become more representative of real factory conditions where processes repeat in predictable loops.

Core claim

The central claim is that introducing a dataset designed with cyclic dynamics arising from the repetitive nature of discrete automation processes fills key gaps, allowing selected MTSAD methods to be evaluated under the federated learning paradigm on both the proposed dataset and a public benchmark dataset.

What carries the argument

A dataset incorporating cyclic dynamics from repetitive discrete automation processes, serving as the testbed for federated MTSAD method evaluation.

If this is right

MTSAD methods can now be tested at larger scale with accurate labels while respecting data privacy constraints of federated learning.
The influence of cyclic process behavior on anomaly detection accuracy becomes directly measurable.
Future method comparisons gain a controlled setting that reflects repetitive industrial automation without the documented flaws of prior data.
Benchmark results on the new dataset provide a baseline for assessing how well methods handle the combination of multivariate series and federated constraints.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The dataset could serve as a template for creating similar resources in other repetitive-process domains such as packaging or assembly lines.
Researchers might test whether adding explicit cycle-length features further boosts detection rates on this data.
The work implies that privacy-preserving federated training may become standard once representative industrial datasets exist.

Load-bearing premise

That existing datasets cannot simultaneously deliver sufficient scale, accurate labels, and freedom from common flaws for federated MTSAD benchmarking.

What would settle it

If evaluations of the selected MTSAD methods on the new cyclic dataset show no measurable difference in detection performance or benchmarking reliability compared with the public dataset that can be traced to the added scale, labels, or cyclic structure.

Figures

Figures reproduced from arXiv: 2605.27486 by Khayyam Nosrati, Martin Uray, Olaf Sassnick, Saverio Messineo, Stefan Huber.

**Figure 1.** Figure 1: This snippet captures around 200 seconds of recorded data. Red highlights a misposition anomaly at a picking position affecting both yaw and pitch variables. requiring a single optimization phase are preferred, excluding approaches with additional optimization during evaluation (e.g., MAD-GAN [9]). The selection further considers inference latency via forward-pass time complexity, problem scope, and featur… view at source ↗

**Figure 2.** Figure 2: Relative performance with respect to CL, where a ratio of 1.0 denotes parity. 6 Discussion The empirical results indicate that architectural heterogeneity and the adoption of state-of-the-art models, such as Transformers and GNNs, do not inherently lead to improved anomaly detection performance. This suggests that increased model capacity and architectural sophistication may introduce unnecessary modeling… view at source ↗

read the original abstract

Federated learning (FL) has broadened the horizon for multivariate time series anomaly detection (MTSAD). However, benchmarking such anomaly detection methods within FL paradigm poses data-centric challenges. The existing datasets do not counteract these challenges since they do not simultaneously provide sufficient scale, accurate labels, and freedom from common flaws. In addition, the role of cyclic process behavior, which is common in discrete industrial automation, remains underexplored for MTSAD for the current state of research. This paper aims to shed more light on the literature and address these gaps by introducing a dataset designed with cyclic dynamics arising from the repetitive nature of discrete automation processes and evaluates selected MTSAD methods on both the proposed dataset and a public benchmark dataset.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper announces a new MTSAD dataset for federated learning in industrial automation but the abstract asserts prior datasets' shortcomings without any comparisons or evidence.

read the letter

The one thing to know is that this is a dataset introduction paper claiming to fix gaps in federated multivariate time series anomaly detection for industrial settings by adding cyclic dynamics from repetitive processes. It positions the new data as better because existing sets supposedly lack scale, accurate labels, and freedom from flaws all at once.

What the paper does is identify a practical need in privacy-sensitive manufacturing where cyclic behavior matters and says it will evaluate some MTSAD methods on the new set plus a public benchmark. That focus on discrete automation cycles is a reasonable angle if the data delivers.

The soft spot is central and not minor. The abstract states the deficiencies of prior datasets as fact but gives no table, numbers, or side-by-side check against SWaT, WADI, MSL, or SMAP on scale, label quality, or common flaws. The stress-test note flags exactly this: without that grounding the motivation and any later results rest on an unverified premise. No results, metrics, or dataset properties appear either, so the claims stay unsupported on the available text.

Since only the abstract was accessible here, it is possible the full manuscript has the required comparison section. On what is visible, though, the load-bearing step does not hold up.

This is for people working on federated anomaly detection who need industrial-style benchmarks with cyclic patterns. A reader building or testing FL models in manufacturing could get value from the dataset if it is released with clear documentation and the missing comparisons.

I would not send it for peer review in its current form because the core motivation lacks evidence. If the full paper supplies the explicit quantification of prior datasets' properties and some validation, then it could deserve referee time for the dataset itself.

Referee Report

1 major / 0 minor

Summary. The paper claims that existing MTSAD datasets fail to simultaneously offer sufficient scale, accurate labels, and freedom from common flaws, while cyclic dynamics from discrete automation remain underexplored in FL settings. It introduces a new dataset incorporating these cyclic properties and evaluates selected MTSAD methods on both the proposed dataset and a public benchmark.

Significance. If the new dataset demonstrably supplies the claimed scale/label/flaw-free combination and the evaluations produce reproducible insights into FL-MTSAD performance under cyclic conditions, the work could strengthen benchmarking practices for industrial anomaly detection. The focus on cyclic process behavior targets a genuine gap if properly validated against prior datasets.

major comments (1)

[Abstract] Abstract (first paragraph): The assertion that 'the existing datasets do not simultaneously provide sufficient scale, accurate labels, and freedom from common flaws' is stated as fact without any supporting comparison (e.g., a table quantifying scale, label quality, and documented flaws for SWaT, WADI, MSL, SMAP or similar). This premise is load-bearing for the motivation and dataset introduction; absent an explicit, citable section with such quantification, the central claim rests on an unverified premise.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for highlighting the need for explicit support of the dataset motivation. We will address this by adding a comparative analysis in the revised manuscript.

read point-by-point responses

Referee: [Abstract] Abstract (first paragraph): The assertion that 'the existing datasets do not simultaneously provide sufficient scale, accurate labels, and freedom from common flaws' is stated as fact without any supporting comparison (e.g., a table quantifying scale, label quality, and documented flaws for SWaT, WADI, MSL, SMAP or similar). This premise is load-bearing for the motivation and dataset introduction; absent an explicit, citable section with such quantification, the central claim rests on an unverified premise.

Authors: We agree that the claim requires explicit quantification to be fully substantiated. In the revised manuscript we will add a table (or dedicated subsection) in the introduction or related-work section that compares SWaT, WADI, MSL, SMAP and similar datasets on the three axes mentioned—scale (sample count, feature dimensionality, duration), label quality (source and verification method), and documented flaws (e.g., sensor drift, missing segments, label noise)—with direct citations to the original dataset papers and known limitations reported in the literature. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical dataset paper with no derivations or fitted quantities

full rationale

The paper introduces a new MTSAD dataset with cyclic dynamics and performs empirical evaluations on it plus a public benchmark. No equations, parameters, or derivations are present in the provided text. The assertion that prior datasets lack scale/labels/flaw-freedom is a factual claim (potentially unsubstantiated) but does not constitute any of the enumerated circularity patterns, as there is no derivation chain that reduces to its own inputs by construction, no self-citation load-bearing on a theorem, and no renaming or ansatz smuggling. This matches the default expectation of no significant circularity for a purely empirical contribution.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No mathematical model or derivation; contribution rests on the empirical claim that existing datasets are deficient and that the new dataset fills the gap. No free parameters, axioms, or invented entities are invoked.

pith-pipeline@v0.9.1-grok · 5657 in / 1112 out tokens · 39730 ms · 2026-06-29T18:42:36.312281+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

32 extracted references · 17 canonical work pages · 1 internal anchor

[1]

In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

Abdulaal, A., Liu, Z., Lancewicki, T.: Practical approach to asynchronous mul- tivariate time series anomaly detection and localization. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. pp. 2485–2494. KDD ’21, Association for Computing Machinery, New York, NY, USA (2021).https://doi.org/10.1145/3447548.3467174

work page doi:10.1145/3447548.3467174 2021
[2]

Sensors24(17) (2024).https:// doi.org/10.3390/s24175492

Alshede, H., Jambi, K., Nassef, L., Alowidi, N., Fadel, E.: Fedavg-p: Performance- based hierarchical federated learning-based anomaly detection system aggregation strategy for advanced metering infrastructure. Sensors24(17) (2024).https:// doi.org/10.3390/s24175492

work page doi:10.3390/s24175492 2024
[3]

In: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

Audibert, J., Michiardi, P., Guyard, F., Marti, S., Zuluaga, M.A.: USAD: Unsu- pervised anomaly detection on multivariate time series. In: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. pp. 3395–3404. KDD ’20, Association for Computing Machinery, New York, NY, USA (2020).https://doi.org/10.1145/3394486.3403392

work page doi:10.1145/3394486.3403392 2020
[4]

IEEE Transactions on Neural Networks and Learning Systems33(6), 2508–2517 (2022).https://doi.org/10

Garg,A.,Zhang,W.,Samaran,J.,Savitha,R.,Foo,C.S.:Anevaluationofanomaly detection and diagnosis in multivariate time series. IEEE Transactions on Neural Networks and Learning Systems33(6), 2508–2517 (2022).https://doi.org/10. 1109/TNNLS.2021.3105827 FL for MTSAD in Industrial Automation 11

work page arXiv 2022
[5]

Deep Learning Scaling is Predictable, Empirically

Hestness, J., Narang, S., Ardalani, N., Diamos, G., Jun, H., Kianinejad, H., Pat- wary, M.M.A., Yang, Y., Zhou, Y.: Deep learning scaling is predictable, empirically. arXiv preprint arXiv:1712.00409 (2017)

work page internal anchor Pith review Pith/arXiv arXiv 2017
[6]

Hundman,K.,Constantinou,V.,Laporte,C.,Colwell,I.,Soderstrom,T.:Detecting spacecraftanomaliesusinglstmsandnonparametricdynamicthresholding.In:Pro- ceedings of the 24th ACM SIGKDD International Conference on Knowledge Dis- covery & Data Mining. pp. 387–395. KDD ’18, Association for Computing Machin- ery, New York, NY, USA (2018).https://doi.org/10.1145/3219...

work page doi:10.1145/3219819.3219845 2018
[7]

Computers in Industry 132, 103509 (2021).https://doi.org/10.1016/j.compind.2021.103509

Huong, T.T., Bac, T.P., Long, D.M., Luong, T.D., Dan, N.M., Quang, L.A., Cong, L.T., Thang, B.D., Tran, K.P.: Detecting cyberattacks using anomaly detection in industrial control systems: A federated learning approach. Computers in Industry 132, 103509 (2021).https://doi.org/10.1016/j.compind.2021.103509

work page doi:10.1016/j.compind.2021.103509 2021
[8]

Jacob, V., Song, F., Stiegler, A., Rad, B., Diao, Y., Tatbul, N.: Exathlon: a bench- mark for explainable anomaly detection over time series. Proc. VLDB Endow. 14(11), 2613–2626 (jul 2021).https://doi.org/10.14778/3476249.3476307

work page doi:10.14778/3476249.3476307 2021
[9]

In: Interna- tional conference on artificial neural networks

Li, D., Chen, D., Jin, B., Shi, L., Goh, J., Ng, S.K.: Mad-gan: Multivariate anomaly detection for time series data with generative adversarial networks. In: Interna- tional conference on artificial neural networks. pp. 703–716. Springer (2019)

2019
[10]

In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining

Li, Z., Zhao, Y., Han, J., Su, Y., Jiao, R., Wen, X., Pei, D.: Multivariate time series anomaly detection and interpretation using hierarchical inter-metric and temporal embedding. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. pp. 3220–3230. KDD ’21, Association for Computing Machinery, New York, NY, USA (2021)

2021
[11]

In: ICC 2020-2020 IEEE international conference on communications (ICC)

Liu, L., Zhang, J., Song, S., Letaief, K.B.: Client-edge-cloud hierarchical federated learning. In: ICC 2020-2020 IEEE international conference on communications (ICC). pp. 1–6. IEEE (2020)

2020
[12]

In: Globerson, A., Mackey, L., Belgrave, D., Fan, A., Paquet, U., Tomczak, J., Zhang, C

Liu, Q., Paparrizos, J.: The elephant in the room: Towards a reliable time-series anomaly detection benchmark. In: Globerson, A., Mackey, L., Belgrave, D., Fan, A., Paquet, U., Tomczak, J., Zhang, C. (eds.) Advances in Neural Information Processing Systems. vol. 37, pp. 108231–108261. Curran Associates, Inc. (2024)

2024
[13]

In: Artificial intelligence and statistics

McMahan, B., Moore, E., Ramage, D., Hampson, S., y Arcas, B.A.: Communication-efficient learning of deep networks from decentralized data. In: Artificial intelligence and statistics. pp. 1273–1282. Pmlr (2017)

2017
[14]

In: 2018 26th Signal Processing and Commu- nications Applications Conference (SIU)

Mirza, A.H., Cosan, S.: Computer network intrusion detection using sequential lstm neural networks autoencoders. In: 2018 26th Signal Processing and Commu- nications Applications Conference (SIU). pp. 1–4 (2018)

2018
[15]

IEEE Access7, 1991– 2005 (2019)

Munir, M., Siddiqui, S.A., Dengel, A., Ahmed, S.: DeepAnT: A deep learning approach for unsupervised anomaly detection in time series. IEEE Access7, 1991– 2005 (2019)

1991
[16]

14590990

Nosrati, K., Uray, M., Messineo, S., Sassnick, O., Huber, S.: Quanser aero 2 pick-and-place dataset (qappd) (May 2026).https://doi.org/10.5281/zenodo. 20287835

work page doi:10.5281/zenodo 2026
[17]

Paparrizos, J., Boniol, P., Palpanas, T., Tsay, R.S., Elmore, A., Franklin, M.J.: Vol- ume under the surface: a new accuracy evaluation measure for time-series anomaly detection. Proc. VLDB Endow.15(11), 2774–2787 (jul 2022)

2022
[18]

IEEE Transactions on Machine Learning in Communications and Networking3, 176–194 (2025)

Raeiszadeh, M., Ebrahimzadeh, A., Glitho, R.H., Eker, J., Mini, R.A.: Asyn- chronous real-time federated learning for anomaly detection in microservice cloud applications. IEEE Transactions on Machine Learning in Communications and Networking3, 176–194 (2025)

2025
[19]

SN computer science2(6), 1–20 (2021) 12 K

Sarker, I.H.: Deep learning: a comprehensive overview on techniques, taxonomy, applications and research directions. SN computer science2(6), 1–20 (2021) 12 K. Nosrati et al

2021
[20]

In: Computer Aided Systems The- ory – EUROCAST 2024

Saßnick, O., Schäfer, G., Rosenstatter, T., Huber, S.: A generative model based honeypot for industrial opc ua communication. In: Computer Aided Systems The- ory – EUROCAST 2024. pp. 320–334. Springer Nature Switzerland, Cham (04 2025).https://doi.org/10.1007/978-3-031-83885-9_29

work page doi:10.1007/978-3-031-83885-9_29 2024
[21]

Data Mining and Knowledge Discovery 38(3), 1027–1068 (2024)

Sørbø, S., Ruocco, M.: Navigating the metric maze: a taxonomy of evaluation met- rics for anomaly detection in time series. Data Mining and Knowledge Discovery 38(3), 1027–1068 (2024)

2024
[22]

NIST Special Publication (SP) NIST SP 800-82r3, National Institute of Standards and Technology, Gaithersburg, MD (2023).https://doi.org/10

Stouffer, K., Pease, M., Tang, C.Y., Zimmerman, T., Pillitteri, V., Lightman, S., Hahn, A., Saravia, S., Sherule, A., Thompson, M.: Guide to operational technology (ot) security. NIST Special Publication (SP) NIST SP 800-82r3, National Institute of Standards and Technology, Gaithersburg, MD (2023).https://doi.org/10. 6028/NIST.SP.800-82r3

2023
[23]

In: Proceed- ings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

Su, Y., Zhao, Y., Niu, C., Liu, R., Sun, W., Pei, D.: Robust anomaly detection for multivariate time series through stochastic recurrent neural network. In: Proceed- ings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. pp. 2828–2837. KDD ’19, Association for Computing Machinery, New York, NY, USA (2019).https://doi.or...

work page doi:10.1145/3292500.3330672 2019
[24]

Electronics11(16) (2022).https://doi.org/10.3390/electronics11162627

Sun, X., Tang, Z., Du, M., Deng, C., Lin, W., Chen, J., Qi, Q., Zheng, H.: A hierarchical federated learning-based intrusion detection system for 5g smart grids. Electronics11(16) (2022).https://doi.org/10.3390/electronics11162627

work page doi:10.3390/electronics11162627 2022
[25]

Computers in Industry 140, 103692 (2022).https://doi.org/10.1016/j.compind.2022.103692

Truong, H.T., Ta, B.P., Le, Q.A., Nguyen, D.M., Le, C.T., Nguyen, H.X., Do, H.T., Nguyen, H.T., Tran, K.P.: Light-weight federated learning-based anomaly detection for time-series data in industrial control systems. Computers in Industry 140, 103692 (2022).https://doi.org/10.1016/j.compind.2022.103692

work page doi:10.1016/j.compind.2022.103692 2022
[26]

Tuli,S.,Casale,G.,Jennings,N.R.:Tranad:deeptransformernetworksforanomaly detection in multivariate time series data. Proc. VLDB Endow.15(6), 1201–1214 (feb 2022).https://doi.org/10.14778/3514061.3514067

work page doi:10.14778/3514061.3514067 2022
[27]

Wagner, D., Michels, T., Schulz, F.C.F., Nair, A., Rudolph, M.R., Kloft, M.: Time- sead: Benchmarking deep multivariate time-series anomaly detection. Trans. Mach. Learn. Res.2023(2023)

2023
[28]

Security and Communication Networks2022(1), 2913293 (2022).https: //doi.org/10.1155/2022/2913293

Wang, K., Li, J., Wu, W.: An efficient intrusion detection method based on fed- erated transfer learning and an extreme learning machine with privacy preser- vation. Security and Communication Networks2022(1), 2913293 (2022).https: //doi.org/10.1155/2022/2913293

work page doi:10.1155/2022/2913293 2022
[29]

In: 2022 IEEE 38th International Conference on Data Engineering (ICDE)

Wu, R., Keogh, E.J.: Current time series anomaly detection benchmarks are flawed and are creating the illusion of progress (extended abstract). In: 2022 IEEE 38th International Conference on Data Engineering (ICDE). pp. 1479–1480 (2022)

2022
[30]

arXiv preprint arXiv:2110.02642 (2021)

Xu, J., Wu, H., Wang, J., Long, M.: Anomaly transformer: Time series anomaly detection with association discrepancy. arXiv preprint arXiv:2110.02642 (2021)

work page arXiv 2021
[31]

In: 2020 IEEE international conference on data mining (ICDM)

Zhao, H., Wang, Y., Duan, J., Huang, C., Cao, D., Tong, Y., Xu, B., Bai, J., Tong,J.,Zhang,Q.:Multivariatetime-seriesanomalydetectionviagraphattention network. In: 2020 IEEE international conference on data mining (ICDM). pp. 841–

2020
[32]

In: 2022 IEEE International Conference on Big Data (Big Data)

Zhu, W., Song, D., Chen, Y., Cheng, W., Zong, B., Mizoguchi, T., Lumezanu, C., Chen, H., Luo, J.: Deep federated anomaly detection for multivariate time series data. In: 2022 IEEE International Conference on Big Data (Big Data). pp. 1–10 (2022).https://doi.org/10.1109/BigData55660.2022.10064694 FL for MTSAD in Industrial Automation 13 Appendix A Detail on...

work page doi:10.1109/bigdata55660.2022.10064694 2022

[1] [1]

In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

Abdulaal, A., Liu, Z., Lancewicki, T.: Practical approach to asynchronous mul- tivariate time series anomaly detection and localization. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. pp. 2485–2494. KDD ’21, Association for Computing Machinery, New York, NY, USA (2021).https://doi.org/10.1145/3447548.3467174

work page doi:10.1145/3447548.3467174 2021

[2] [2]

Sensors24(17) (2024).https:// doi.org/10.3390/s24175492

Alshede, H., Jambi, K., Nassef, L., Alowidi, N., Fadel, E.: Fedavg-p: Performance- based hierarchical federated learning-based anomaly detection system aggregation strategy for advanced metering infrastructure. Sensors24(17) (2024).https:// doi.org/10.3390/s24175492

work page doi:10.3390/s24175492 2024

[3] [3]

In: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

Audibert, J., Michiardi, P., Guyard, F., Marti, S., Zuluaga, M.A.: USAD: Unsu- pervised anomaly detection on multivariate time series. In: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. pp. 3395–3404. KDD ’20, Association for Computing Machinery, New York, NY, USA (2020).https://doi.org/10.1145/3394486.3403392

work page doi:10.1145/3394486.3403392 2020

[4] [4]

IEEE Transactions on Neural Networks and Learning Systems33(6), 2508–2517 (2022).https://doi.org/10

Garg,A.,Zhang,W.,Samaran,J.,Savitha,R.,Foo,C.S.:Anevaluationofanomaly detection and diagnosis in multivariate time series. IEEE Transactions on Neural Networks and Learning Systems33(6), 2508–2517 (2022).https://doi.org/10. 1109/TNNLS.2021.3105827 FL for MTSAD in Industrial Automation 11

work page arXiv 2022

[5] [5]

Deep Learning Scaling is Predictable, Empirically

Hestness, J., Narang, S., Ardalani, N., Diamos, G., Jun, H., Kianinejad, H., Pat- wary, M.M.A., Yang, Y., Zhou, Y.: Deep learning scaling is predictable, empirically. arXiv preprint arXiv:1712.00409 (2017)

work page internal anchor Pith review Pith/arXiv arXiv 2017

[6] [6]

Hundman,K.,Constantinou,V.,Laporte,C.,Colwell,I.,Soderstrom,T.:Detecting spacecraftanomaliesusinglstmsandnonparametricdynamicthresholding.In:Pro- ceedings of the 24th ACM SIGKDD International Conference on Knowledge Dis- covery & Data Mining. pp. 387–395. KDD ’18, Association for Computing Machin- ery, New York, NY, USA (2018).https://doi.org/10.1145/3219...

work page doi:10.1145/3219819.3219845 2018

[7] [7]

Computers in Industry 132, 103509 (2021).https://doi.org/10.1016/j.compind.2021.103509

Huong, T.T., Bac, T.P., Long, D.M., Luong, T.D., Dan, N.M., Quang, L.A., Cong, L.T., Thang, B.D., Tran, K.P.: Detecting cyberattacks using anomaly detection in industrial control systems: A federated learning approach. Computers in Industry 132, 103509 (2021).https://doi.org/10.1016/j.compind.2021.103509

work page doi:10.1016/j.compind.2021.103509 2021

[8] [8]

Jacob, V., Song, F., Stiegler, A., Rad, B., Diao, Y., Tatbul, N.: Exathlon: a bench- mark for explainable anomaly detection over time series. Proc. VLDB Endow. 14(11), 2613–2626 (jul 2021).https://doi.org/10.14778/3476249.3476307

work page doi:10.14778/3476249.3476307 2021

[9] [9]

In: Interna- tional conference on artificial neural networks

Li, D., Chen, D., Jin, B., Shi, L., Goh, J., Ng, S.K.: Mad-gan: Multivariate anomaly detection for time series data with generative adversarial networks. In: Interna- tional conference on artificial neural networks. pp. 703–716. Springer (2019)

2019

[10] [10]

In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining

Li, Z., Zhao, Y., Han, J., Su, Y., Jiao, R., Wen, X., Pei, D.: Multivariate time series anomaly detection and interpretation using hierarchical inter-metric and temporal embedding. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. pp. 3220–3230. KDD ’21, Association for Computing Machinery, New York, NY, USA (2021)

2021

[11] [11]

In: ICC 2020-2020 IEEE international conference on communications (ICC)

Liu, L., Zhang, J., Song, S., Letaief, K.B.: Client-edge-cloud hierarchical federated learning. In: ICC 2020-2020 IEEE international conference on communications (ICC). pp. 1–6. IEEE (2020)

2020

[12] [12]

In: Globerson, A., Mackey, L., Belgrave, D., Fan, A., Paquet, U., Tomczak, J., Zhang, C

Liu, Q., Paparrizos, J.: The elephant in the room: Towards a reliable time-series anomaly detection benchmark. In: Globerson, A., Mackey, L., Belgrave, D., Fan, A., Paquet, U., Tomczak, J., Zhang, C. (eds.) Advances in Neural Information Processing Systems. vol. 37, pp. 108231–108261. Curran Associates, Inc. (2024)

2024

[13] [13]

In: Artificial intelligence and statistics

McMahan, B., Moore, E., Ramage, D., Hampson, S., y Arcas, B.A.: Communication-efficient learning of deep networks from decentralized data. In: Artificial intelligence and statistics. pp. 1273–1282. Pmlr (2017)

2017

[14] [14]

In: 2018 26th Signal Processing and Commu- nications Applications Conference (SIU)

Mirza, A.H., Cosan, S.: Computer network intrusion detection using sequential lstm neural networks autoencoders. In: 2018 26th Signal Processing and Commu- nications Applications Conference (SIU). pp. 1–4 (2018)

2018

[15] [15]

IEEE Access7, 1991– 2005 (2019)

Munir, M., Siddiqui, S.A., Dengel, A., Ahmed, S.: DeepAnT: A deep learning approach for unsupervised anomaly detection in time series. IEEE Access7, 1991– 2005 (2019)

1991

[16] [16]

14590990

Nosrati, K., Uray, M., Messineo, S., Sassnick, O., Huber, S.: Quanser aero 2 pick-and-place dataset (qappd) (May 2026).https://doi.org/10.5281/zenodo. 20287835

work page doi:10.5281/zenodo 2026

[17] [17]

Paparrizos, J., Boniol, P., Palpanas, T., Tsay, R.S., Elmore, A., Franklin, M.J.: Vol- ume under the surface: a new accuracy evaluation measure for time-series anomaly detection. Proc. VLDB Endow.15(11), 2774–2787 (jul 2022)

2022

[18] [18]

IEEE Transactions on Machine Learning in Communications and Networking3, 176–194 (2025)

Raeiszadeh, M., Ebrahimzadeh, A., Glitho, R.H., Eker, J., Mini, R.A.: Asyn- chronous real-time federated learning for anomaly detection in microservice cloud applications. IEEE Transactions on Machine Learning in Communications and Networking3, 176–194 (2025)

2025

[19] [19]

SN computer science2(6), 1–20 (2021) 12 K

Sarker, I.H.: Deep learning: a comprehensive overview on techniques, taxonomy, applications and research directions. SN computer science2(6), 1–20 (2021) 12 K. Nosrati et al

2021

[20] [20]

In: Computer Aided Systems The- ory – EUROCAST 2024

Saßnick, O., Schäfer, G., Rosenstatter, T., Huber, S.: A generative model based honeypot for industrial opc ua communication. In: Computer Aided Systems The- ory – EUROCAST 2024. pp. 320–334. Springer Nature Switzerland, Cham (04 2025).https://doi.org/10.1007/978-3-031-83885-9_29

work page doi:10.1007/978-3-031-83885-9_29 2024

[21] [21]

Data Mining and Knowledge Discovery 38(3), 1027–1068 (2024)

Sørbø, S., Ruocco, M.: Navigating the metric maze: a taxonomy of evaluation met- rics for anomaly detection in time series. Data Mining and Knowledge Discovery 38(3), 1027–1068 (2024)

2024

[22] [22]

NIST Special Publication (SP) NIST SP 800-82r3, National Institute of Standards and Technology, Gaithersburg, MD (2023).https://doi.org/10

Stouffer, K., Pease, M., Tang, C.Y., Zimmerman, T., Pillitteri, V., Lightman, S., Hahn, A., Saravia, S., Sherule, A., Thompson, M.: Guide to operational technology (ot) security. NIST Special Publication (SP) NIST SP 800-82r3, National Institute of Standards and Technology, Gaithersburg, MD (2023).https://doi.org/10. 6028/NIST.SP.800-82r3

2023

[23] [23]

In: Proceed- ings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

Su, Y., Zhao, Y., Niu, C., Liu, R., Sun, W., Pei, D.: Robust anomaly detection for multivariate time series through stochastic recurrent neural network. In: Proceed- ings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. pp. 2828–2837. KDD ’19, Association for Computing Machinery, New York, NY, USA (2019).https://doi.or...

work page doi:10.1145/3292500.3330672 2019

[24] [24]

Electronics11(16) (2022).https://doi.org/10.3390/electronics11162627

Sun, X., Tang, Z., Du, M., Deng, C., Lin, W., Chen, J., Qi, Q., Zheng, H.: A hierarchical federated learning-based intrusion detection system for 5g smart grids. Electronics11(16) (2022).https://doi.org/10.3390/electronics11162627

work page doi:10.3390/electronics11162627 2022

[25] [25]

Computers in Industry 140, 103692 (2022).https://doi.org/10.1016/j.compind.2022.103692

Truong, H.T., Ta, B.P., Le, Q.A., Nguyen, D.M., Le, C.T., Nguyen, H.X., Do, H.T., Nguyen, H.T., Tran, K.P.: Light-weight federated learning-based anomaly detection for time-series data in industrial control systems. Computers in Industry 140, 103692 (2022).https://doi.org/10.1016/j.compind.2022.103692

work page doi:10.1016/j.compind.2022.103692 2022

[26] [26]

Tuli,S.,Casale,G.,Jennings,N.R.:Tranad:deeptransformernetworksforanomaly detection in multivariate time series data. Proc. VLDB Endow.15(6), 1201–1214 (feb 2022).https://doi.org/10.14778/3514061.3514067

work page doi:10.14778/3514061.3514067 2022

[27] [27]

Wagner, D., Michels, T., Schulz, F.C.F., Nair, A., Rudolph, M.R., Kloft, M.: Time- sead: Benchmarking deep multivariate time-series anomaly detection. Trans. Mach. Learn. Res.2023(2023)

2023

[28] [28]

Security and Communication Networks2022(1), 2913293 (2022).https: //doi.org/10.1155/2022/2913293

Wang, K., Li, J., Wu, W.: An efficient intrusion detection method based on fed- erated transfer learning and an extreme learning machine with privacy preser- vation. Security and Communication Networks2022(1), 2913293 (2022).https: //doi.org/10.1155/2022/2913293

work page doi:10.1155/2022/2913293 2022

[29] [29]

In: 2022 IEEE 38th International Conference on Data Engineering (ICDE)

Wu, R., Keogh, E.J.: Current time series anomaly detection benchmarks are flawed and are creating the illusion of progress (extended abstract). In: 2022 IEEE 38th International Conference on Data Engineering (ICDE). pp. 1479–1480 (2022)

2022

[30] [30]

arXiv preprint arXiv:2110.02642 (2021)

Xu, J., Wu, H., Wang, J., Long, M.: Anomaly transformer: Time series anomaly detection with association discrepancy. arXiv preprint arXiv:2110.02642 (2021)

work page arXiv 2021

[31] [31]

In: 2020 IEEE international conference on data mining (ICDM)

Zhao, H., Wang, Y., Duan, J., Huang, C., Cao, D., Tong, Y., Xu, B., Bai, J., Tong,J.,Zhang,Q.:Multivariatetime-seriesanomalydetectionviagraphattention network. In: 2020 IEEE international conference on data mining (ICDM). pp. 841–

2020

[32] [32]

In: 2022 IEEE International Conference on Big Data (Big Data)

Zhu, W., Song, D., Chen, Y., Cheng, W., Zong, B., Mizoguchi, T., Lumezanu, C., Chen, H., Luo, J.: Deep federated anomaly detection for multivariate time series data. In: 2022 IEEE International Conference on Big Data (Big Data). pp. 1–10 (2022).https://doi.org/10.1109/BigData55660.2022.10064694 FL for MTSAD in Industrial Automation 13 Appendix A Detail on...

work page doi:10.1109/bigdata55660.2022.10064694 2022