Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

A New Perspective on Time Series Anomaly Detection: Faster Patch-based Broad Learning System

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read A shallow patch-based broad learning system with a contrastive kernel-perturbation branch outperforms twelve deep and classical baselines in unsupervised time series anomaly detection while training in seconds.

desk verdict A clever and fast BLS-based TSAD architecture with an attractive empirical story, but the central KL anomaly score is ill-posed as written and the results lack error bars and code. read the letter →

arxiv 2412.05498 v1 pith:HVTDC63P submitted 2024-12-07 cs.LG cs.AI

classification cs.LGcs.AI
keywords timeseriesanomalydetectionbroadlearningsystempatchingcontrastivemulti-scaleensemblekernelperturbationunsupervisedpseudo-inversetraining
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper is trying to establish that deep learning architectures are not a prerequisite for state-of-the-art unsupervised time series anomaly detection. It proposes CPatchBLS, a shallow Broad Learning System that segments series into patches, runs two branches (one with a random Gaussian-kernel perturbation), and scores anomalies by the symmetric KL divergence between the branches' reconstructions, then averages that score across multiple patch sizes. Across five widely used real-world datasets, the paper reports that this shallow system outperforms twelve deep and classical baselines on ROC-AUC, ROC-PR, and PA-F1, while training in about 12.6 seconds in parallel or 33.6 seconds serially. A careful reader would care because it offers a concrete alternative to deep models in settings where training speed and simplicity matter, such as industrial monitoring.

What carries the argument

The load-bearing object is the Simple Kernel Perturbation (SKP) branch, which applies an approximate Gaussian random feature map $\mathrm{SKP}(Z) = \frac{1}{\sqrt{d_k}}[\sqrt{2}\cos(\omega_1^\top Z + b_1), \dots, \sqrt{2}\cos(\omega_{d_k}^\top Z + b_{d_k})]$ with $\omega_i \sim \mathcal{N}(0,\sigma^2)$ and $b_i \sim U[0,2\pi]$ to the feature layer of one PatchBLS branch. The anomaly score is the symmetrized KL divergence $\mathrm{Scorediff} = \frac{1}{2}\mathrm{KL}(\mathrm{SKP\text{-}branch}, \mathrm{Basic\text{-}branch}) + \frac{1}{2}\mathrm{KL}(\mathrm{Basic\text{-}branch}, \mathrm{SKP\text{-}branch})$, averaged over multiple patch sizes. This random projection creates a second viewpoint of the same patches without backpropagation; the divergence between the two reconstruction views is what separates normal from anomalous windows. Training is done by pseudo-inverse computation of the output weights rather than gradient descent, which is what keeps the whole system fast.

What would settle it

On a fixed benchmark, replace the random Gaussian feature projection (Eq. 11) with a deterministic linear or identity transform of the same output dimension; if ROC-AUC and PA-F1 stay essentially unchanged, the kernel-perturbation branch is not the mechanism carrying the performance. A second test: inject labeled point and pattern anomalies into a normal-only training set and check whether the symmetric KL score (Eq. 13) ranks them above normal windows with AUC appreciably above chance for each anomaly type.

Watch

Extended reading notes

Core claim

The central claim, on the paper's own terms, is that a shallow BLS extended with patching and a contrastive perturbation branch can surpass both deep learning and classical machine learning methods for TSAD. The construction is Dual-PatchBLS: a Basic-PatchBLS branch and an SKP-PatchBLS branch whose feature-layer output is mapped through a random approximation of a Gaussian kernel, with the anomaly score defined as the symmetrized KL divergence between their reconstructions. A Multi-Scale Patches ensemble averages these scores over several patch sizes to reduce sensitivity to any single patch length. The reported result is average ROC-AUC of 99.81%, ROC-PR of 98.25%, and PA-F1 of 96.87% across MSL, SMAP, SWaT, WADI, and PSM, with training time of 12.6 s (parallel) or 33.6 s (serial), which the paper interprets as evidence that complex deep architectures are not required for high-performance anomaly detection.

Load-bearing premise

The whole anomaly signal rests on the assumption that randomly projecting one branch's features through an approximate Gaussian kernel makes normal and anomalous patches diverge under the symmetric KL divergence, so that the resulting score ranks anomalies above normal windows.

Editorial extensions

If this is right

  • If the reported benchmarks are representative, state-of-the-art TSAD does not require deep networks; a shallow, pseudo-inverse-trained system can hold or beat the accuracy bar.
  • Training and testing times of tens of seconds (or about 13 s in parallel) make per-dataset retraining practical in industrial monitoring pipelines where deep models taking minutes to hours are prohibitive.
  • The multi-scale patch ensemble reduces the sensitivity to patch-length choice that a single Dual-PatchBLS shows, so detection is more stable across point and pattern anomalies.
  • The ablation path from PatchBLS to SKP-PatchBLS to Dual-PatchBLS to CPatchBLS shows each added component (nonlinear perturbation, contrastive comparison, multi-scale integration) contributes a measurable improvement, which is a direct corollary of the paper's tables.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension is to treat the SKP branch as a cheap negative view generator and apply the same KL-divergence score to other shallow or random-feature detectors; if the divergence signal is what matters, similar gains should appear without BLS-specific machinery.
  • The symmetric KL between two stochastic reconstructions can be read as a measure of representation instability: anomalies are windows where a random feature perturbation changes the reconstruction disproportionately. This interpretation is not stated in the paper and could be tested by comparing the score to the variance across multiple random seeds.
  • Because BLS supports incremental learning, one could retrain the output layer on sliding windows to adapt to distribution drift while keeping the random features fixed; the paper does not explore this, but the speed figures suggest it is feasible.
  • The comparison covers five standard benchmarks and twelve baselines as of the paper's writing; a natural stress test is a head-to-head on newer 2024 contrastive detectors and on datasets with more subtle, long-range anomalies, where single-patch local semantics may be limiting.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes CPatchBLS, an unsupervised time-series anomaly detection method that combines the Broad Learning System (BLS) with patching, a dual-branch contrastive scheme, and multi-scale ensembling. Time series are segmented into patches, passed through randomly generated feature and enhancement layers, and the output weights are solved by pseudo-inverse. A second branch introduces a Simple Kernel Perturbation (SKP) based on random Fourier features. The anomaly score is the symmetrized KL divergence between the reconstructions of the two branches, averaged over several patch sizes. On five real-world benchmarks (MSL, SMAP, SWaT, WADI, PSM), the paper reports average ROC-AUC 99.81%, ROC-PR 98.25%, and PA-F1 96.87%, with training time 12.6 s (parallel) or 33.6 s (serial), and claims to outperform twelve baselines in both accuracy and speed.

Significance. If the empirical results are reproducible, the paper makes a useful contribution: it demonstrates that a shallow, closed-form-trained model can rival and in many cases surpass deep-learning anomaly detectors while being substantially faster. The manuscript includes pseudo-code, a complexity analysis, and ablation studies that provide partial evidence for the contributions of the SKP branch and multi-scale ensembling. The main limitation is that the central anomaly mechanism is not well-defined as written, and the empirical claims lack variance estimates, so the current evidence is not yet sufficient to support the strong performance claims.

major comments (3)
  1. [Section III-D, Eq. (13)] The anomaly score is defined as a symmetrized KL divergence between the reconstruction outputs of the two branches. KL divergence is only defined for probability distributions, but the reconstructions Y in Eq. (8) are real-valued matrices (N_patch x S_patch), and the manuscript never specifies a normalization, softmax, histogram, or other distributional transformation. As written, the score can be negative, infinite, or undefined, making the central mechanism ill-posed and unreproducible from the published equations. The authors should specify the exact transformation used before computing the KL, and state how the detection threshold δ is determined in practice.
  2. [Section IV-D, Table II] The experimental comparison reports a single run per method with no standard deviations, number of seeds, or code. The BLS weights are randomly generated (Eq. (1)), so run-to-run variance is expected. The claimed improvements over the best baseline are small on average (0.77% in ROC-AUC), and for some datasets/metrics the proposed method is not the best (e.g., SWaT ROC-AUC: AnomTrans 99.95 vs. CPatchBLS 99.90). Without variance estimates or multiple seeds, it is impossible to assess whether the reported differences are statistically significant. The authors should provide multi-seed results and the exact hyperparameter settings for each dataset.
  3. [Section IV-H and Fig. 7] The sensitivity experiments on MSL, SMAP, and PSM are used to select patch lists and hyperparameters, and the same datasets are then used for the final comparison in Table II. This creates a selection-on-the-test-set risk: the reported numbers may reflect the best values achieved during the sensitivity sweep rather than a fixed, pre-specified configuration. To support the claim that CPatchBLS generalizes, the selection should be based on a validation split or a nested procedure, or the authors should explicitly state that the reported results are the best over the sensitivity grid.
minor comments (5)
  1. [Abstract and Section IV-D] The statement that the method 'outperforms previous deep learning and machine learning methods' is too strong; CPatchBLS does not achieve the best ROC-AUC on SWaT (AnomTrans 99.95 vs. 99.90), and IForest has lower training time. Please qualify the claim as 'on average' or 'in most metrics'.
  2. [Section IV-G, conclusions] The second conclusion says 'In most cases (except on the WADI dataset), PatchBLS outperforms SKP-PatchBLS,' but Table III shows the opposite: SKP-PatchBLS is better on MSL, SMAP, SWaT, and PSM, with PatchBLS better only on WADI. This interpretation contradicts the reported numbers and should be corrected.
  3. [Section IV-D] The explanation that CPatchBLS performed particularly well on WADI 'due to the better data quality' is speculative and not supported by any analysis. It should be removed or replaced with a concrete argument.
  4. [Section III-D, Eq. (11)] The dimension of the random vector ω_i is not specified. For random Fourier features, ω_i should have the same dimension as the feature vector Z, and this should be stated explicitly. Also, the bias b_i is described as U[0, 2π] in the text but Eq. (11) uses b_i without a range; please align the notation.
  5. [Throughout] There are numerous typos and notation inconsistencies: 'Technolegy' in the affiliation, 'integratin multi-scalele features' in contribution 3, 'Sequantial' in Section IV-E, and Algorithm 1 uses loop variables m and q while the text defines Gft and Cft. Please proofread and unify the notation.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; the anomaly score is constructed from branch reconstructions fitted only to input patches, not from labels or from the benchmark metrics.

full rationale

The derivation chain is self-contained. PatchBLS (Eqs. 1-10) uses randomly generated feature/enhancement nodes and a pseudo-inverse ridge regression whose target Y is the input patch matrix itself; this is ordinary reconstruction, not a fit to anomaly labels. Eq. (13) defines Scorediff as a symmetrized KL divergence between the reconstructions of the Basic-PatchBLS and SKP-PatchBLS branches, and Eq. (14) averages such scores across patch scales. This is a stated anomaly-score construction, not a prediction derived from the ground-truth labels or from the evaluation metrics. No fitted parameter is renamed as a prediction: the SKP branch weights are random (Eq. 11), and the only fitted weights (Eq. 10) are least-squares reconstruction weights. The paper cites its own earlier works (AdaMemBLS, SIMAD, PatchAD) as motivation and related work, but those citations are not used to justify the correctness of CPatchBLS or to forbid alternative approaches; the reported superiority is tested against twelve external baselines on five public benchmarks. The central mechanism in Eq. (13) is arguably ill-posed because KL divergence is not defined for unnormalized real-valued reconstructions without a stated normalization, but that is a well-formedness and reproducibility concern, not a circularity. Therefore, no load-bearing step reduces to its own inputs.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The method introduces no new physical entities. The main ledger items are the many hyperparameters tuned on the evaluation datasets and the implicit assumptions about reconstruction and divergence as anomaly signals. The largest uncharged cost is the per-dataset patch-list selection, which directly controls the final score.

free parameters (6)
  • Patch-size list per dataset (multi-scale patches) = SMAP: [61, 6, 22]; other datasets not stated
    Final anomaly scores average Dual-PatchBLS models at several patch sizes (Eq. 14); the list is chosen empirically per dataset (Fig. 7) and is not reported for MSL, SWaT, WADI, PSM.
  • Feature-layer dimension, groups, and cascade layers (Dft, Cft, Gft) = Tuned to peaks in [4, 8] groups/cascades; exact final values not given
    Sensitivity experiments (Fig. 6a-c) select these; the final configuration is omitted.
  • Enhancement-layer dimension, groups, and cascade layers (Denh, Cenh, Genh) = Tuned via Fig. 6d-f; exact final values not given
    Depth and width of enhancement nodes affect reconstruction and are tuned on the evaluation datasets.
  • SKP kernel dimension dk and sigma = Tuned via Fig. 6g-h; exact values not stated
    These control the random Fourier projection in Eq. (11) and hence the dual-branch divergence score.
  • Output regularization r and shrink coefficient s = r in [0.1, 0.3], s in [0.8, 1.0]
    Selected by sensitivity analysis (Fig. 6i-j); no final values or validation protocol reported.
  • Detection threshold delta for PA-F1 and decision = Not described
    PA-F1 requires a threshold or point-adjustment procedure, but the paper does not specify how delta is chosen.
assumptions (6)
  • standard math The pseudo-inverse update Wo = (A^T A + lambda I)^{-1} A^T Y (Eq. 10) minimizes the regularized least-squares objective (Eq. 9).
    Standard linear algebra; unproved in the paper but accepted background.
  • standard math Random Fourier features of the form in Eq. (11) provide an approximation to a Gaussian kernel.
    Used to justify SKP-PatchBLS; the paper does not prove the approximation error or its effect on anomaly scoring.
  • domain assumption Multivariate time series can be processed channel-independently, one univariate series at a time.
    Section III-C adopts PatchTST's channel independence; cross-channel dependencies are ignored.
  • domain assumption Reconstruction error and the discrepancy between the two branches indicate anomalies.
    The anomaly score is the KL divergence between Basic-PatchBLS and SKP-PatchBLS reconstructions (Eq. 13); no proof connects this divergence to anomaly likelihood.
  • domain assumption Training data predominantly contains normal patterns and the model can represent them.
    Problem statement in Section III-A says the model is trained on Xtrain expected to represent normal patterns; contaminated training data would break the assumption.
  • domain assumption Randomly initialized feature and enhancement nodes, after sparse autoencoder and orthogonalization, form a useful representation for reconstruction.
    This is inherited from BLS literature (ref 16) and assumed without guarantees in this setting.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A New Perspective on Time Series Anomaly Detection: Faster Patch-based Broad Learning System." pith.science (2026). https://pith.science/paper/HVTDC63P

@misc{pith2026241205498,
  author       = {Pith},
  title        = {Pith review of: A New Perspective on Time Series Anomaly Detection: Faster Patch-based Broad Learning System},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HVTDC63P}},
  note         = {Machine review of arXiv:2412.05498}
}
read the original abstract

Time series anomaly detection (TSAD) has been a research hotspot in both academia and industry in recent years. Deep learning methods have become the mainstream research direction due to their excellent performance. However, new viewpoints have emerged in recent TSAD research. Deep learning is not required for TSAD due to limitations such as slow deep learning speed. The Broad Learning System (BLS) is a shallow network framework that benefits from its ease of optimization and speed. It has been shown to outperform machine learning approaches while remaining competitive with deep learning. Based on the current situation of TSAD, we propose the Contrastive Patch-based Broad Learning System (CPatchBLS). This is a new exploration of patching technique and BLS, providing a new perspective for TSAD. We construct Dual-PatchBLS as a base through patching and Simple Kernel Perturbation (SKP) and utilize contrastive learning to capture the differences between normal and abnormal data under different representations. To compensate for the temporal semantic loss caused by various patching, we propose CPatchBLS with model level integration, which takes advantage of BLS's fast feature to build model-level integration and improve model detection. Using five real-world series anomaly detection datasets, we confirmed the method's efficacy, outperforming previous deep learning and machine learning methods while retaining a high level of computing efficiency.

Figures

Figures reproduced from arXiv: 2412.05498 by the authors.

Figure 1
Figure 1. By analyzing key metrics like ROC-AUC, PR-AUC, and PA-F1 for twelve state-of-the-art models and our methods on five datasets, the figure shows [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Overview of CPatchBLS. First, the Patching Module segments multivariate data into univariate time series. Second, SKP-PatchBLS alongside Basic [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The overview of PatchBLS. original representation. The specific equation of SKP is as follows: SKP(Z) = 1 √ dk [ √ 2 cos(ω ⊤ 1 Z + b1), . . . , √ 2 cos(ω ⊤ dk Z + bdk )]. (11) where dk denotes the dimension of the random feature space, the variable Z represents time series data following processing by the feature layer, wi ∼ N (0, σ2 ) represents the Gaussian random vector, and bi ∼ U[0, 2π] represents the random va… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Comparison of average time consumption of machine learning methods [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Average Training and Testing Time Comparison Results on MSL, SMAP, SWAT, WADI, and PSM Datasets. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Sensitivity analysis of the proposed method on MSL, SMAP, and PSM datasets. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Sensitivity analysis w.r.t patch lists of the proposed method on SMAP [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. CCE: Confidence-Consistency Evaluation for Time Series Anomaly Detection

    cs.LG 2025-09 conditional novelty 6.0 of 10

    CCE is a bounded, robust, O(n) event-level metric for time series anomaly detection that combines prediction confidence with Bayesian uncertainty consistency, and RankEval is a new benchmark for comparing such metrics.

Reference graph

Works this paper leans on

40 extracted references · 24 canonical work pages · cited by 1 Pith paper

  1. [1]

    Adap- tive memory broad learning system for unsupervised time series anomaly detection,

    Z. Zhong, Z. Yu, Z. Fan, C. P. Chen, and K. Yang, “Adap- tive memory broad learning system for unsupervised time series anomaly detection,” IEEE Transactions on Neural Networks and Learning Systems , 2024

  2. [2]

    Anomaly trans- former: Time series anomaly detection with association discrepancy,

    J. Xu, H. Wu, J. Wang, and M. Long, “Anomaly trans- former: Time series anomaly detection with association discrepancy,” arXiv preprint arXiv:2110.02642 , 2021

  3. [3]

    Anomalybert: Self-supervised transformer for time se- ries anomaly detection using data degradation scheme,

    Y . Jeong, E. Yang, J. H. Ryu, I. Park, and M. Kang, “Anomalybert: Self-supervised transformer for time se- ries anomaly detection using data degradation scheme,” arXiv preprint arXiv:2305.04468 , 2023

  4. [4]

    Dcde- tector: Dual attention contrastive representation learning for time series anomaly detection,

    Y . Yang, C. Zhang, T. Zhou, Q. Wen, and L. Sun, “Dcde- tector: Dual attention contrastive representation learning for time series anomaly detection,” in Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2023, pp. 3033–3045

  5. [5]

    Insightful simplicity: Dissimilarity in time series anomaly detection,

    Z. Zhong, Z. Yu, J. Chen, and K. Yang, “Insightful simplicity: Dissimilarity in time series anomaly detection,” in Proceedings of the ACM Turing Award Celebration Conference - China 2024 , ser. ACM-TURC ’24. New York, NY , USA: Association for Computing Machinery, 2024, p. 242–243. [Online]. Available: https://doi.org/10.1145/3674399.3674486

  6. [6]

    Simad: A simple dissimilarity-based approach for time series anomaly detection,

    Z. Zhong, Z. Yu, X. Xi, Y . Xu, J. Chen, and K. Yang, “Simad: A simple dissimilarity-based approach for time series anomaly detection,” 05 2024

  7. [7]

    Zhong, Z

    Z. Zhong, Z. Yu, Y . Yang, W. Wang, and K. Yang, “Patchad: A lightweight patch-based mlp-mixer 5 10 15 20 (a) the dimension of feature nodes 80.0 82.5 85.0 87.5 90.0 92.5 95.0 97.5 100.0PA-F1 Score % MSL SMAP PSM 2 4 6 8 10 (b) the group numbers of feature nodes 80.0 82.5 85.0 87.5 90.0 92.5 95.0 97.5 100.0PA-F1 Score % MSL SMAP PSM 2 4 6 8 10 (c) the num...

  8. [8]

    [22] [61] [22,6] [61,6] [61,22] [61,6,22] (k) the different patches lists 65 70 75 80 85 90 95 100 Score (%) 96.63% 99.64% 99.83% 99.80% 99.88% 99.86% 99.86% 76.89% 97.23% 98.42% 98.41% 99.13% 99.20% 99.23% 69.58% 93.38% 96.43% 95.88% 96.57% 96.35% 97.35% ROC-AUC ROC-PR F1-PA Fig. 7. Sensitivity analysis w.r.t patch lists of the proposed method on SMAP da...

Show all 40 references
  1. [9]

    Patchmixer: A patch- mixing architecture for long-term time series forecast- ing,

    Z. Gong, Y . Tang, and J. Liang, “Patchmixer: A patch- mixing architecture for long-term time series forecast- ing,” arXiv preprint arXiv:2310.00655 , 2023

  2. [10]

    Are transform- ers effective for time series forecasting?

    A. Zeng, M. Chen, L. Zhang, and Q. Xu, “Are transform- ers effective for time series forecasting?” in Proceedings of the AAAI conference on artificial intelligence , vol. 37, no. 9, 2023, pp. 11 121–11 128

  3. [11]

    Position: Quo vadis, unsupervised time series anomaly detection?

    M. S. Sarfraz, M.-Y . Chen, L. Layer, K. Peng, and M. Koulakis, “Position: Quo vadis, unsupervised time series anomaly detection?” in Forty-first International Conference on Machine Learning

  4. [12]

    Estimating the support of a high- dimensional distribution,

    B. Sch ¨olkopf, J. C. Platt, J. Shawe-Taylor, A. J. Smola, and R. C. Williamson, “Estimating the support of a high- dimensional distribution,” Neural computation , vol. 13, no. 7, pp. 1443–1471, 2001

  5. [13]

    Isolation forest,

    F. T. Liu, K. M. Ting, and Z.-H. Zhou, “Isolation forest,” in 2008 eighth ieee international conference on data mining. IEEE, 2008, pp. 413–422

  6. [14]

    Lof: identifying density-based local outliers,

    M. M. Breunig, H.-P. Kriegel, R. T. Ng, and J. Sander, “Lof: identifying density-based local outliers,” in Pro- ceedings of the 2000 ACM SIGMOD international con- ference on Management of data , 2000, pp. 93–104

  7. [15]

    Efficient algorithms for mining outliers from large data sets,

    S. Ramaswamy, R. Rastogi, and K. Shim, “Efficient algorithms for mining outliers from large data sets,” in Proceedings of the 2000 ACM SIGMOD international conference on Management of data , 2000, pp. 427–438

  8. [16]

    Deep learning for anomaly detection in time-series data: Review, anal- ysis, and guidelines,

    K. Choi, J. Yi, C. Park, and S. Yoon, “Deep learning for anomaly detection in time-series data: Review, anal- ysis, and guidelines,” IEEE access, vol. 9, pp. 120 043– 120 065, 2021

  9. [17]

    Broad learning system: An effec- tive and efficient incremental learning system without the need for deep architecture,

    C. P. Chen and Z. Liu, “Broad learning system: An effec- tive and efficient incremental learning system without the need for deep architecture,” IEEE transactions on neural networks and learning systems, vol. 29, no. 1, pp. 10–24, 2017

  10. [18]

    A time series is worth 64 words: Long-term forecasting with transformers,

    Y . Nie, N. H. Nguyen, P. Sinthong, and J. Kalagnanam, “A time series is worth 64 words: Long-term forecasting with transformers,” arXiv preprint arXiv:2211.14730 , 2022

  11. [19]

    Learning to em- bed time series patches independently,

    S. Lee, T. Park, and K. Lee, “Learning to em- bed time series patches independently,” arXiv preprint arXiv:2312.16427, 2023

  12. [20]

    Sparse coding with an overcomplete basis set: A strategy employed by v1?

    B. A. Olshausen and D. J. Field, “Sparse coding with an overcomplete basis set: A strategy employed by v1?” Vision research, vol. 37, no. 23, pp. 3311–3325, 1997

  13. [21]

    Re- search of broad learning system,

    C. Ren, C. Yuan, Y . Sun, Z. Liu, and C. P. Chen, “Re- search of broad learning system,” Application Research of Computers/Jisuanji Yingyong Yanjiu , vol. 38, no. 8, 2021

  14. [22]

    Sensitivehue: Multivariate time series anomaly detection by enhancing the sensitivity to normal patterns,

    Y . Feng, W. Zhang, Y . Fu, W. Jiang, J. Zhu, and W. Ren, “Sensitivehue: Multivariate time series anomaly detection by enhancing the sensitivity to normal patterns,” in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2024, pp. 782– 793

  15. [23]

    Detecting spacecraft anomalies us- ing lstms and nonparametric dynamic thresholding,

    K. Hundman, V . Constantinou, C. Laporte, I. Colwell, and T. Soderstrom, “Detecting spacecraft anomalies us- ing lstms and nonparametric dynamic thresholding,” in Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, 2018, pp. 387–395

  16. [24]

    The soil moisture active passive (smap) mission,

    D. Entekhabi, E. G. Njoku, P. E. O’neill, K. H. Kellogg, W. T. Crow, W. N. Edelstein, J. K. Entin, S. D. Goodman, T. J. Jackson, J. Johnson et al. , “The soil moisture active passive (smap) mission,” Proceedings of the IEEE, vol. 98, no. 5, pp. 704–716, 2010

  17. [25]

    A dataset to support research in the design of secure water treatment systems,

    J. Goh, S. Adepu, K. N. Junejo, and A. Mathur, “A dataset to support research in the design of secure water treatment systems,” in Critical Information Infrastruc- tures Security: 11th International Conference, CRITIS 2016, Paris, France, October 10–12, 2016, Revised Se- lecte...

  18. [26]

    Wadi: a water distribution testbed for research in the design of secure cyber physical systems,

    C. M. Ahmed, V . R. Palleti, and A. P. Mathur, “Wadi: a water distribution testbed for research in the design of secure cyber physical systems,” in Proceedings of the 3rd international workshop on cyber-physical systems for smart water networks , 2017, pp. 25–28

  19. [27]

    Practical ap- proach to asynchronous multivariate time series anomaly detection and localization,

    A. Abdulaal, Z. Liu, and T. Lancewicki, “Practical ap- proach to asynchronous multivariate time series anomaly detection and localization,” in Proceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining, 2021, pp. 2485–2494

  20. [28]

    Daemon: Unsupervised anomaly detection and interpretation for multivariate time series,

    X. Chen, L. Deng, F. Huang, C. Zhang, Z. Zhang, Y . Zhao, and K. Zheng, “Daemon: Unsupervised anomaly detection and interpretation for multivariate time series,” in 2021 IEEE 37th International Conference on Data Engineering (ICDE). IEEE, 2021, pp. 2225–2230

  21. [29]

    Tranad: Deep transformer networks for anomaly detection in multivari- ate time series data,

    S. Tuli, G. Casale, and N. R. Jennings, “Tranad: Deep transformer networks for anomaly detection in multivari- ate time series data,” arXiv preprint arXiv:2201.07284 , 2022

  22. [30]

    Learning multi-pattern normal- ities in the frequency domain for efficient time series anomaly detection,

    F. Chen, Y . Zhang, Z. Qin, L. Fan, R. Jiang, Y . Liang, Q. Wen, and S. Deng, “Learning multi-pattern normal- ities in the frequency domain for efficient time series anomaly detection,” in 2024 IEEE 40th International Conference on Data Engineering (ICDE) . IEEE, 2024, pp. 747–760

  23. [31]

    Temporal-frequency masked autoencoders for time se- ries anomaly detection,

    Y . Fang, J. Xie, Y . Zhao, L. Chen, Y . Gao, and K. Zheng, “Temporal-frequency masked autoencoders for time se- ries anomaly detection,” in 2024 IEEE 40th International Conference on Data Engineering (ICDE) . IEEE, 2024, pp. 1228–1241

  24. [32]

    Ecod: Unsupervised outlier detection using empirical cumulative distribution functions,

    Z. Li, Y . Zhao, X. Hu, N. Botta, C. Ionescu, and G. H. Chen, “Ecod: Unsupervised outlier detection using empirical cumulative distribution functions,”IEEE Trans- actions on Knowledge and Data Engineering , vol. 35, no. 12, pp. 12 181–12 193, 2022

  25. [33]

    Deep one-class classification,

    L. Ruff, R. Vandermeulen, N. Goernitz, L. Deecke, S. A. Siddiqui, A. Binder, E. M ¨uller, and M. Kloft, “Deep one-class classification,” in International conference on machine learning. PMLR, 2018, pp. 4393–4402

  26. [34]

    Usad: Unsupervised anomaly detection on multivariate time series,

    J. Audibert, P. Michiardi, F. Guyard, S. Marti, and M. A. Zuluaga, “Usad: Unsupervised anomaly detection on multivariate time series,” in Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining , 2020, pp. 3395–3404

  27. [35]

    Fits: Modeling time series with 10k parameters,

    Z. Xu, A. Zeng, and Q. Xu, “Fits: Modeling time series with 10k parameters,” in The Twelfth International Conference on Learning Representations

  28. [36]

    Timesnet: Temporal 2d-variation modeling for general time series analysis,

    H. Wu, T. Hu, Y . Liu, H. Zhou, J. Wang, and M. Long, “Timesnet: Temporal 2d-variation modeling for general time series analysis,” arXiv preprint arXiv:2210.02186 , 2022

  29. [37]

    One fits all: Power general time series analysis by pretrained lm,

    T. Zhou, P. Niu, L. Sun, R. Jin et al., “One fits all: Power general time series analysis by pretrained lm,” Advances in neural information processing systems , vol. 36, pp. 43 322–43 355, 2023

  30. [38]

    Calibrated one-class classification for unsupervised time series anomaly detection,

    H. Xu, Y . Wang, S. Jian, Q. Liao, Y . Wang, and G. Pang, “Calibrated one-class classification for unsupervised time series anomaly detection,” IEEE Transactions on Knowl- edge and Data Engineering , 2024

  31. [39]

    Long short term memory networks for anomaly detection in time series

    P. Malhotra, L. Vig, G. Shroff, P. Agarwal et al., “Long short term memory networks for anomaly detection in time series.” in Esann, vol. 2015, 2015, p. 89

  32. [40]

    The elephant in the room: Towards a reliable time-series anomaly detection bench- mark,

    Q. Liu and J. Paparrizos, “The elephant in the room: Towards a reliable time-series anomaly detection bench- mark,” in NeurIPS 2024, 2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.