Pith. sign in

REVIEW 5 major objections 6 minor 40 references

NeuCoReClass AD: Redefining Self-Supervised Time Series Anomaly Detection

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read NeuCoReClass AD claims that training a time series model on contrastive, reconstruction, and classification tasks together, with learned neural transformations, yields better anomaly detection and unsupervised anomaly-type characterization.

desk verdict The multi-task claim is undercut by the paper's own ablation; the anomaly-characterization idea is the real novelty but needs better validation. read the letter →

arxiv 2508.00909 v2 pith:4JX4YBJA submitted 2025-07-29 cs.LG

Aitor Sánchez-Ferrera , Usue Mori , Borja Calvo , Jose A. Lozano This is my paper · ORCID
classification cs.LG
keywords timeseriesanomalydetectionself-supervisedlearningcontrastivereconstructionneuraltransformationmulti-taskunsupervisedcharacterizationone-classclassification
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

NeuCoReClass AD sets out to show that the choice of proxy task is not a zero-sum game: a single encoder trained on contrastive, reconstruction, and classification objectives together, using neural transformations learned from normal data instead of handcrafted domain-specific ones, gives better time series anomaly detection. The paper reports that across 144 benchmark problems this multi-task model beats classical baselines and most deep-learning alternatives, with its clearest advantage under the harder (N-1)-vs-rest protocol where most classes are normal. It also claims the per-transformation parts of the anomaly score form a descriptor that separates distinct anomaly types without any labels.

What carries the argument

The load-bearing mechanism is the joint loss in Eq. (6), which combines three terms: a contrastive loss (Eq. 2) that pulls together representations of views generated by the same learned neural transformation and pushes apart views from different transformations; a reconstruction loss (Eq. 4) that decodes each augmented view back to the original sample; and a cross-entropy classification loss (Eq. 5) that predicts which transformation generated each view. The three terms are combined with learned uncertainty weights, and at test time the same three objectives define per-task anomaly scores that are summed into the final score. The transformation-wise contributions of that score act as a descriptor vector for each sample, which is what enables the unsupervised characterization of anomaly types.

What would settle it

Train NeuCoReClass AD with the full batch contrastive loss, then compare, on held-out normal and anomalous samples, the ranking produced by the single-sample score of Eq. (7) with the ranking produced by the actual batch loss when that sample is inserted into a large normal minibatch; if the two rankings disagree materially, the per-sample contrastive score is not faithful to the training objective.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that the three proxy tasks are complementary in one framework: the contrastive loss organizes the latent space by transformation while preventing collapse, the reconstruction loss stops the latent codes from losing the original signal, and the classification loss sharpens transformation-specific subspaces. The paper shows that a single network trained with all three objectives detects anomalies competitively under one-vs-rest evaluation and with the best average AUROC and AUPR under the harder (N-1)-vs-rest setting, while models trained with any single task or pair of tasks are measurably worse. It also demonstrates that the vector of per-transformation anomaly contributions separates anomaly classes in dimensionality-reduced projections, enabling unsupervised characterization of anomaly types.

Load-bearing premise

The per-sample contrastive anomaly score in Eq. (7) assumes that summing negative similarities among one sample's own augmented views reproduces the separation that the batch-level contrastive loss enforced during training; if that assumption fails, the contrastive part of the score is measuring something other than what the training objective optimized.

Editorial extensions

If this is right

  • Anomaly detectors no longer need domain-specific handcrafted transformations, because the transformations are learned from normal data by the model itself.
  • Combining the three proxy tasks is preferable to single-task or two-task self-supervised training, since the ablation results show the full model matches the best pair and adds the contrastive signal needed for anomaly-type separation.
  • Under the more difficult (N-1)-vs-rest protocol, where most classes are normal, the method reports the best mean AUROC and AUPR among all compared methods.
  • The same architecture provides anomaly detection and anomaly profiling in one pass, because the per-transformation contribution vector exposes structure in the anomaly space.
  • Performance is stable across the tested choices of the number of transformations K and the contrastive temperature parameter.

Reading between the lines

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

  • Beyond the paper, the per-transformation contribution vectors could be clustered to produce weak labels for downstream supervised training, effectively turning an unsupervised detector into a label generator.
  • Beyond the paper, the anomaly-type separation is demonstrated on one dataset, so a natural check is whether the same separation appears on other multivariate benchmarks; if it does, transformation contributions become a general interpretability channel for time series outliers.
  • Beyond the paper, the single-sample contrastive score of Eq. (7) can be directly compared against a minibatch-based score on held-out data; a large disagreement would indicate that a different per-sample surrogate is needed for the contrastive component.
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

5 major / 6 minor

Summary. The paper proposes NeuCoReClass AD, a self-supervised multi-task framework for time series anomaly detection that jointly trains contrastive, reconstruction, and classification proxy tasks with learned neural transformations. The authors evaluate on 144 problems derived from 20 UCR datasets under one-vs-rest and (N-1)-vs-rest protocols, comparing against shallow baselines (OCSVM, IF, LOF) and deep baselines (AutoAD, DAGMM, DeepSVDD, FixedTS, NeuTraL AD). They report mean AUROC/AUPR and average ranks over five seeds, finding that NeuCoReClass AD is competitive with DeepSVDD in one-vs-rest and superior in (N-1)-vs-rest. The paper additionally claims that per-transformation anomaly contributions enable unsupervised characterization of anomaly types, illustrated with t-SNE on the Epilepsy dataset. An ablation study in Appendix C reports that the full model performs essentially the same as a variant using only reconstruction+classification, and that removing the contrastive loss weakens the qualitative anomaly-characterization visualization.

Significance. If the central claims held, the contribution would be a general, transformation-agnostic self-supervised TSAD method with a useful anomaly-profiling capability. The paper ships public code, evaluates across a broad suite of 144 problems with both univariate and multivariate data, and follows the common NeuTraL AD evaluation protocol, which aids comparability. The anomaly score is a direct combination of training losses with no test-set fitting, which is a methodological strength. However, the significance is currently constrained by the ablation result showing no detection benefit from the contrastive task and by the qualitative-only evidence for the characterization claim; the headline 'multi-task superiority' is therefore not established as stated.

major comments (5)
  1. [Appendix C, Tables VIII and IX] The ablations show that Reconstruction+Classification alone matches or slightly beats the full model: under one-vs-rest, AUROC 73.18 vs 73.22 and AUPR 81.50 vs 81.42; under (N-1)-vs-rest, AUROC 66.85 vs 66.63 and AUPR 47.10 vs 46.77. Since the central claim of the paper is that integrating contrastive, reconstruction, and classification proxy tasks yields consistently superior anomaly detection, this result directly undermines that claim. The paper needs either a statistically significant detection gain from the contrastive task or a re-framing of the contribution to detection as coming primarily from reconstruction+classification, with the contrastive task justified separately.
  2. [Section II-D, Eq. (7)] The per-sample contrastive anomaly score AScon is presented without derivation, and the indices in Eq. (7) are undefined: x_k^i and x_q^j appear with no specification of i or j for a single new sample. Unlike Eq. (2), which sums over a minibatch, Eq. (7) replaces the positive-sum term by 1 and the negative sum with a single-sample expression. It is not shown that this score approximates the training contrastive objective, and the notation makes it impossible to verify. Given that the contrastive term also contributes no measurable detection improvement in the ablations, this score needs either a clear derivation and correct notation or removal from the anomaly-score definition.
  3. [Section IV-A, Tables I and II, and Section III-D] All quantitative comparisons are reported as point estimates averaged over five seeds, with no error bars, confidence intervals, or paired significance tests. Under one-vs-rest, DeepSVDD actually exceeds NeuCoReClass AD by +0.40 AUROC and +0.23 AUPR, yet the abstract states the method 'consistently outperforms both classical baselines and most deep-learning alternatives' and the text calls the DeepSVDD margin 'negligible' without statistical support. The authors should report standard deviations or paired tests (e.g., Wilcoxon signed-rank) across the 144 problems; the claim of consistent superiority is not supported by the current evidence.
  4. [Section IV-C and Appendix C, Figures 4 and 5] The anomaly-characterization claim rests entirely on t-SNE visualizations of a single dataset (Epilepsy). No formal measure of 'transformation-wise anomaly contribution' is defined, despite this being the object being visualized, and no quantitative clustering or separation metric (e.g., adjusted Rand index, silhouette score, or class-conditional distances) is reported. The claim that the contrastive loss is 'essential' for characterization is supported only by a visual comparison between two t-SNE plots. This should be substantiated with a precise definition of the contribution vector and quantitative evaluation, ideally on more than one dataset.
  5. [Section III-A and Appendix A] The benchmark selection is restricted to the first 10 UCR univariate and first 10 UCR multivariate datasets satisfying simple inclusion criteria, with no justification that these are representative of time series anomaly detection tasks more broadly. While this follows prior work, the 'wide range of benchmarks' claim in the abstract is stronger than the evidence: all datasets are classification archives converted to anomaly detection, and no datasets from the common TSAD literature (e.g., SMD, SWaT, SMAP) are included. The authors should either temper the generalization claim or extend the evaluation.
minor comments (6)
  1. [Abstract] There are typos: 'accross' should be 'across', and 'tailored specific domains' should be 'tailored to specific domains'.
  2. [Table I] The column header 'Método' is Spanish; it should be 'Method'.
  3. [Section II-C] The update descriptions contain garbled notation: 'Tθk , θφ, θψ' and 'Tθk , θφ, θf ' should be written as separate parameter sets, e.g., θ_{T_k}, θ_φ, θ_ψ, to be readable.
  4. [Section II-C.2] 'This is enfoced by the reconstruction loss' should be 'enforced', and the sentence beginning 'In the backward pass, This loss' has an incorrect capitalization.
  5. [Section IV-A] The description of Table I says 'NeuCoReClass AD's entries are shaded in dark gray whenever it is not the top-performing method,' but no shading is visible in the provided tables; this formatting note should either be removed or implemented.
  6. [Appendix B] The results tables would benefit from showing standard deviations across the five seeds rather than only the means, especially since seed variability is used to claim robustness.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: anomaly scores are direct training losses, evaluation is against external benchmarks, and self-citations are only background.

full rationale

NeuCoReClass AD's anomaly score is built directly from the three training losses rather than from any quantity fitted to the test set: AS_rec equals Eq. (4), AS_class equals Eq. (5), AS_con is the modified contrastive loss in Eq. (7), and the final score in Eq. (8) is a weighted sum using the same uncertainty parameters learned during training. No parameter is fitted to the evaluation set, so there is no 'fitted input called prediction'. The claimed superiority is tested on UCR benchmark problems against external baselines (IF, LOF, OC-SVM, AutoAD, DAGMM, DeepSVDD, FixedTS, NeuTraL AD), not derived from an assumption that builds in the conclusion. Self-citations such as [2], [6], and [8] appear only in background passages (definitions, review of proxy-task conditions, and an example of fixed transformations); the load-bearing design choices and comparisons are supported by external references ([19], [24], [25], [27], [37]) and by the paper's own experiments. Eq. (7) has undefined indices (i, j) and the per-transformation contribution vector in Sec. IV-C is not formally defined, but these are rigor gaps, not circular reductions. Appendix C honestly reports that Reconstruction+Classification alone matches or slightly beats the full model; this weakens the multi-task-superiority claim empirically but does not make the derivation circular. The characterization claim is demonstrated only on one dataset via t-SNE, which is weak evidence, but again a validation-strength issue rather than a self-referential one. Overall, the derivation chain does not reduce any claimed result to its own inputs.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The method relies on standard unsupervised AD assumptions, a heuristic per-sample contrastive score, and a narrow benchmark selection. The learned uncertainty weights and hyperparameters are the main fitted quantities.

free parameters (4)
  • learned uncertainty weights sigma_con, sigma_rec, sigma_class = learned during training (values not reported)
    Equation 6; they weight the three losses and the final anomaly score. They are fit to normal training data and directly affect AS in Eq. 8.
  • number of transformations K = 12
    Section III-C.2: fixed to 12 for all methods; sensitivity shown on two datasets.
  • contrastive temperature tau = 0.1
    Section III-C.3: fixed to 0.1 following prior work; sensitivity on two datasets.
  • batch size = 32 (16 for two datasets)
    Section III-C.2; set to accommodate memory, but affects contrastive positives.
assumptions (5)
  • domain assumption Training data contains only normal samples (or negligible anomalies).
    Standard unsupervised AD assumption; Section II-C uses normal samples only.
  • domain assumption The three proxy task criteria (disruption, diversity, preservation) capture normality.
    Section II-B states these rules are derived from prior work [6], [19], [24]; they are heuristic.
  • ad hoc to paper Identity transformation T1 is included and treated as a class.
    Section II-A.1; justified by citing [6], but it affects classification and contrastive objectives.
  • ad hoc to paper The per-sample contrastive score in Eq. 7 approximates the batch contrastive loss.
    Section II-D; no derivation, undefined indices.
  • domain assumption The first 10 univariate and 10 multivariate UCR datasets satisfying the criteria are representative.
    Section III-A; no standard TSAD benchmarks.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NeuCoReClass AD: Redefining Self-Supervised Time Series Anomaly Detection." pith.science (2026). https://pith.science/paper/4JX4YBJA

@misc{pith2026250800909,
  author       = {Pith},
  title        = {Pith review of: NeuCoReClass AD: Redefining Self-Supervised Time Series Anomaly Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4JX4YBJA}},
  note         = {Machine review of arXiv:2508.00909}
}
read the original abstract

Time series anomaly detection plays a critical role in a wide range of real-world applications. Among unsupervised approaches, self-supervised learning has gained traction for modeling normal behavior without the need of labeled data. However, many existing methods rely on a single proxy task, limiting their ability to capture meaningful patterns in normal data. Moreover, they often depend on handcrafted transformations tailored specific domains, hindering their generalization accross diverse problems. To address these limitations, we introduce NeuCoReClass AD, a self-supervised multi-task time series anomaly detection framework that combines contrastive, reconstruction, and classification proxy tasks. Our method employs neural transformation learning to generate augmented views that are informative, diverse, and coherent, without requiring domain-specific knowledge. We evaluate NeuCoReClass AD across a wide range of benchmarks, demonstrating that it consistently outperforms both classical baselines and most deep-learning alternatives. Furthermore, it enables the characterization of distinct anomaly profiles in a fully unsupervised manner.

Figures

Figures reproduced from arXiv: 2508.00909 by the authors.

Figure 1
Figure 1. In Section II-A, we describe the architecture of NeuCoRe￾Class AD. Following, Section II-B defines the desired proper￾ties that the generated augmented views should satisfy to be effective across the proxy tasks considered in our approach. Section II-C details the training process of NeuCoReClass AD. Finally, II-D presents how the representations of the augmented views of new samples can be used to compute their ano… view at source ↗
Figure 1
Figure 1. Overview of NeuCoReClass AD. The top panel shows the generation process of the augmented views and their [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Visualization of the latent space induced by the joint training with reconstruction, contrastive, and classification [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figures from the paper (7 more)
Figure 3
Figure 3. Figure 3: Comparison of the reconstruction errors per input dimension between normal and anomalous test samples. Different [PITH_FULL_IMAGE:figures/full_fig_p009_3.png]
Figure 4
Figure 4. Figure 4: Class separability in the Epilepsy dataset across different feature spaces. Each point corresponds to a sample from the evaluation set, and each color indicates one of the ground-truth classes. [10] X. Liu, F. Zhang, Z. Hou, L. Mian, Z. Wang, J. Zhang, and J. Tang, “Se…
Figure 5
Figure 5. Figure 5: Class separability in the Epilepsy dataset across different feature spaces for the ablated model. Each point corresponds to a sample from the evaluation set, and each color indicates one of the ground-truth classes. 3 4 7 11 15 Number of transformations (K) 0 20 40 60 …
Figure 6
Figure 6. Figure 6: Sensitivity of NeuCoReClass AD to the number of transformations [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: Sensitivity of NeuCoReClass AD to the number of transformations [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: Sensitivity of NeuCoReClass AD to the temperature parameter [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]
Figure 9
Figure 9. Figure 9: Sensitivity of NeuCoReClass AD to the temperature parameter [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 26 canonical work pages

  1. [1]

    J. D. Hamilton, Time series analysis. Princeton university press, 1994

    work page 1994

  2. [2]

    Analyzing rare event, anomaly, novelty and outlier detection terms under the supervised classification framework,

    A. Carre ˜no, I. Inza, and J. A. Lozano, “Analyzing rare event, anomaly, novelty and outlier detection terms under the supervised classification framework,” Artificial Intelligence Review, vol. 53, pp. 3575–3594, 2020

    work page 2020

  3. [3]

    Anomaly detection for iot time- series data: A survey,

    A. A. Cook, G. Mısırlı, and Z. Fan, “Anomaly detection for iot time- series data: A survey,” IEEE Internet of Things Journal , vol. 7, no. 7, pp. 6481–6494, 2019

    2019

  4. [4]

    Financial fraud: a review of anomaly detection techniques and recent advances,

    W. Hilal, S. A. Gadsden, and J. Yawney, “Financial fraud: a review of anomaly detection techniques and recent advances,” Expert systems With applications, vol. 193, p. 116429, 2022

    work page 2022

  5. [5]

    Learning representations from healthcare time series data for unsupervised anomaly detection,

    J. Pereira and M. Silveira, “Learning representations from healthcare time series data for unsupervised anomaly detection,” in 2019 IEEE international conference on big data and smart computing (BigComp) . IEEE, 2019, pp. 1–7

    work page 2019

  6. [6]

    A review on self- supervised learning for time series anomaly detection: Recent advances and open challenges,

    A. S ´anchez-Ferrera, B. Calvo, and J. A. Lozano, “A review on self- supervised learning for time series anomaly detection: Recent advances and open challenges,” arXiv preprint arXiv:2501.15196 , 2025

    work page arXiv 2025

  7. [7]

    Anomaly detection: A survey,

    V . Chandola, A. Banerjee, and V . Kumar, “Anomaly detection: A survey,” ACM computing surveys (CSUR) , vol. 41, no. 3, pp. 1–58, 2009

    work page 2009

  8. [8]

    Water leak detection using self-supervised time series classification,

    A. Bl ´azquez-Garc´ıa, A. Conde, U. Mori, and J. A. Lozano, “Water leak detection using self-supervised time series classification,” Information Sciences, vol. 574, pp. 528–541, 2021

    work page 2021

Show all 40 references
  1. [9]

    Deep learning for time series anomaly detection: A survey,

    Z. Zamanzadeh Darban, G. I. Webb, S. Pan, C. Aggarwal, and M. Salehi, “Deep learning for time series anomaly detection: A survey,” ACM Computing Surveys, vol. 57, no. 1, pp. 1–42, 2024. 10 Real Class epilepsy running sawing walking (a) Dimensionality reduction over the origina...

    2024

  2. [10]

    Self-supervised learning: Generative or contrastive,

    X. Liu, F. Zhang, Z. Hou, L. Mian, Z. Wang, J. Zhang, and J. Tang, “Self-supervised learning: Generative or contrastive,” IEEE transactions on knowledge and data engineering , vol. 35, no. 1, pp. 857–876, 2021

    2021

  3. [11]

    Self-supervised anomaly detection: A survey and outlook,

    H. Hojjati, T. K. K. Ho, and N. Armanfard, “Self-supervised anomaly detection: A survey and outlook,” arXiv preprint arXiv:2205.05173 , 2022

    2022 arXiv

  4. [12]

    Autoencoders,

    D. Bank, N. Koenigstein, and R. Giryes, “Autoencoders,” arXiv preprint arXiv:2003.05991, 2020

    2003 arXiv

  5. [13]

    Mad: Self-supervised masked anomaly detection task for multivariate time series,

    Y . Fu and F. Xue, “Mad: Self-supervised masked anomaly detection task for multivariate time series,” in 2022 International Joint Conference on Neural Networks (IJCNN) . IEEE, 2022, pp. 1–8

    2022

  6. [14]

    Deepfib: Self-imputation for time series anomaly detection,

    M. Liu, Z. Xu, and Q. Xu, “Deepfib: Self-imputation for time series anomaly detection,” arXiv preprint arXiv:2112.06247 , 2021

    2021 arXiv

  7. [15]

    Anomaly detection using autoencoders with nonlinear dimensionality reduction,

    M. Sakurada and T. Yairi, “Anomaly detection using autoencoders with nonlinear dimensionality reduction,” in Proceedings of the MLSDA 2014 2nd workshop on machine learning for sensory data analysis , 2014, pp. 4–11

    2014

  8. [16]

    A novel approach for automatic acoustic novelty detection using a denoising autoencoder with bidirectional lstm neural networks,

    E. Marchi, F. Vesperini, F. Eyben, S. Squartini, and B. Schuller, “A novel approach for automatic acoustic novelty detection using a denoising autoencoder with bidirectional lstm neural networks,” in 2015 IEEE international conference on acoustics, speech and signal processing...

    2015

  9. [17]

    Self-supervised anomaly detection via neural autore- gressive flows with active learning,

    J. Zhang, K. Saleeby, T. Feldhausen, S. Bi, A. Plotkowski, and D. Womble, “Self-supervised anomaly detection via neural autore- gressive flows with active learning,” Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States), Tech. Rep., 2021

    2021

  10. [18]

    Task-oriented self-supervised learning for anomaly detection in elec- troencephalography,

    Y . Zheng, Z. Liu, R. Mo, Z. Chen, W.-s. Zheng, and R. Wang, “Task-oriented self-supervised learning for anomaly detection in elec- troencephalography,” in International Conference on Medical Image Computing and Computer-Assisted Intervention . Springer, 2022, pp. 193–203

    2022

  11. [19]

    Neural transformation learning for deep anomaly detection beyond images,

    C. Qiu, T. Pfrommer, M. Kloft, S. Mandt, and M. Rudolph, “Neural transformation learning for deep anomaly detection beyond images,” in International conference on machine learning. PMLR, 2021, pp. 8703– 8714

    2021

  12. [20]

    A survey on contrastive self-supervised learning,

    A. Jaiswal, A. R. Babu, M. Z. Zadeh, D. Banerjee, and F. Makedon, “A survey on contrastive self-supervised learning,” Technologies, vol. 9, no. 1, p. 2, 2020

    2020

  13. [21]

    Time-series anomaly detec- tion via contextual discriminative contrastive learning,

    K. Chen, M. Feng, and T. S. Wirjanto, “Time-series anomaly detec- tion via contextual discriminative contrastive learning,” arXiv preprint arXiv:2304.07898, 2023

    2023 arXiv

  14. [22]

    Detecting anomalies within time series using local neural transformations,

    T. Schneider, C. Qiu, M. Kloft, D. A. Latif, S. Staab, S. Mandt, and M. Rudolph, “Detecting anomalies within time series using local neural transformations,” arXiv preprint arXiv:2202.03944 , 2022

    2022 arXiv

  15. [23]

    Multi-task self-supervised visual learn- ing,

    C. Doersch and A. Zisserman, “Multi-task self-supervised visual learn- ing,” in Proceedings of the IEEE international conference on computer vision, 2017, pp. 2051–2060

    2017

  16. [24]

    Data augmentation is a hyperparameter: Cherry-picked self-supervision for unsupervised anomaly detection is creating the illusion of success,

    J. Yoo, T. Zhao, and L. Akoglu, “Data augmentation is a hyperparameter: Cherry-picked self-supervision for unsupervised anomaly detection is creating the illusion of success,” arXiv preprint arXiv:2208.07734, 2022

    2022 arXiv

  17. [25]

    A simple framework for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” in International conference on machine learning . PMLR, 2020, pp. 1597–1607

    2020

  18. [26]

    Deep contrastive one-class time series anomaly detection,

    R. Wang, C. Liu, X. Mou, K. Gao, X. Guo, P. Liu, T. Wo, and X. Liu, “Deep contrastive one-class time series anomaly detection,” in Proceedings of the 2023 SIAM International Conference on Data Mining (SDM). SIAM, 2023, pp. 694–702

    2023

  19. [27]

    Auxiliary tasks in multi-task learning,

    L. Liebel and M. K ¨orner, “Auxiliary tasks in multi-task learning,” arXiv preprint arXiv:1805.06334, 2018

    2018 arXiv

  20. [28]

    The ucr time series archive,

    H. A. Dau, A. Bagnall, K. Kamgar, C.-C. M. Yeh, Y . Zhu, S. Gharghabi, C. A. Ratanamahatana, and E. Keogh, “The ucr time series archive,” IEEE/CAA Journal of Automatica Sinica , vol. 6, no. 6, pp. 1293–1305, 2019

    2019

  21. [29]

    Support vector method for novelty detection,

    B. Sch ¨olkopf, R. C. Williamson, A. Smola, J. Shawe-Taylor, and J. Platt, “Support vector method for novelty detection,” Advances in neural information processing systems , vol. 12, 1999

    1999

  22. [30]

    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

    2008

  23. [31]

    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 Proceedings of the 2000 ACM SIGMOD international conference on Management of data , 2000, pp. 93–104

    2000

  24. [32]

    Autoencoder-based network anomaly detection,

    Z. Chen, C. K. Yeo, B. S. Lee, and C. T. Lau, “Autoencoder-based network anomaly detection,” in 2018 Wireless telecommunications sym- posium (WTS). IEEE, 2018, pp. 1–5

    2018

  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

    2018

  26. [34]

    Deep autoencoding gaussian mixture model for unsupervised anomaly detection,

    B. Zong, Q. Song, M. R. Min, W. Cheng, C. Lumezanu, D. Cho, and H. Chen, “Deep autoencoding gaussian mixture model for unsupervised anomaly detection,” in International conference on learning representa- tions, 2018

    2018

  27. [35]

    Effective end-to-end unsupervised outlier detection via inlier priority of discriminative network,

    S. Wang, Y . Zeng, X. Liu, E. Zhu, J. Yin, C. Xu, and M. Kloft, “Effective end-to-end unsupervised outlier detection via inlier priority of discriminative network,” Advances in neural information processing systems, vol. 32, 2019

    2019

  28. [36]

    Scikit-learn: Machine learning in python,

    F. Pedregosa, G. Varoquaux, A. Gramfort, V . Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V . Dubourg et al. , “Scikit-learn: Machine learning in python,” the Journal of machine Learning research, vol. 12, pp. 2825–2830, 2011

    2011

  29. [37]

    Ts2vec: Towards universal representation of time series,

    Z. Yue, Y . Wang, J. Duan, T. Yang, C. Huang, Y . Tong, and B. Xu, “Ts2vec: Towards universal representation of time series,” in Proceed- ings of the AAAI conference on artificial intelligence , vol. 36, no. 8, 2022, pp. 8980–8987

    2022

  30. [38]

    Effi- 11 cient time series anomaly detection by multiresolution self-supervised discriminative network,

    D. Huang, L. Shen, Z. Yu, Z. Zheng, M. Huang, and Q. Ma, “Effi- 11 cient time series anomaly detection by multiresolution self-supervised discriminative network,” Neurocomputing, vol. 491, pp. 261–272, 2022

    2022

  31. [39]

    Visualizing data using t-sne

    L. Van der Maaten and G. Hinton, “Visualizing data using t-sne.” Journal of machine learning research , vol. 9, no. 11, 2008

    2008

  32. [40]

    On hyperparameter tuning in general clustering problemsm,

    X. Fan, Y . Yue, P. Sarkar, and Y . R. Wang, “On hyperparameter tuning in general clustering problemsm,” in International conference on machine learning. PMLR, 2020, pp. 2996–3007. 12 APPENDIX A DATASETS We summarize the key characteristics of each dataset used in our experime...

    2020

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.