Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Noise-Resilient Point-wise Anomaly Detection in Time Series Using Weak Segment Labels

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Point-level anomalies in time series can be localized from coarse, noisy segment-level labels by combining positive-unlabeled learning with a temporal smoothness and separability loss, and the resulting framework reports the best F1 among…

desk verdict A practical PU-learning pipeline for point-level TSAD from weak segment labels, with broad experiments but a central formal gap: the per-point score h(X) is undefined and conflicts with the pooling in Eq. (2). read the letter →

arxiv 2501.11959 v1 pith:LTMK3ZBW submitted 2025-01-21 cs.LG

Yaxuan Wang , Hao Cheng , Jing Xiong , Qingsong Wen , Han Jia , Ruixuan Song , Liyuan Zhang , Zhaowei Zhu
show 1 more author
Yang Liu
This is my paper · ORCID
classification cs.LG
keywords timeseriesanomalydetectionweaklysupervisedlearningpositiveandunlabelednoisylabelssegment-levelpoint-leveltemporalconstraintlossmultivariate
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

This paper tries to establish that point-level anomalies in multivariate time series can be detected from weak, noisy segment-level labels—labels that say an anomalous event occurred somewhere in a window, but not where, with most windows entirely unlabeled. It proposes NRdetector, a two-stage framework: a coarse PU-learning stage that trains a segment classifier on labeled positives plus confidently selected negatives, regularized by a loss encouraging smooth per-point anomaly scores and separation between positive and negative segments; then a fine-grained stage that ranks points inside predicted positive segments and thresholds them with an estimated anomaly rate. The authors report that NRdetector beats 13 weakly, semi-, and unsupervised baselines under both pure F1 and PA%K F1 on all five real-world datasets (EMG, SMD, PSM, MSL, SMAP), with the largest margins at high label-noise rates. If true, the practical payoff is that cheap event-level annotation, not expensive point-level annotation, could be enough to localize anomalies.

What carries the argument

The load-bearing machinery is the training objective, not a single architectural novelty. Stage-1 uses the non-negative PU risk estimator $R_{pu}$ to classify segments, together with the Time-Constraint loss $\mathcal{L}_c = \lambda_1\mathcal{L}_{\text{smooth}} + \lambda_2\mathcal{L}_{\text{sep}}$: $\mathcal{L}_{\text{smooth}}$ penalizes squared differences between anomaly scores of adjacent points inside a segment, and $\mathcal{L}_{\text{sep}}$ pushes the mean score of unlabeled (treated-negative) segments below that of labeled positives. Before training, a sample selector builds a Katz-index similarity graph over segments, extracts reliable negatives far from labeled positives, and propagates labels to clean the unlabeled set. Stage-2 then treats each predicted positive segment as a bag of $T$ points, sorts their per-point anomaly scores, and uses a clusterability-based anomaly-rate estimator to set the cutoff automatically. The paper argues that the PU risk upper-bounds the true classification risk and that reducing label noise tightens the generalization bound, which is what makes the pipeline noise-resilient.

What would settle it

Open the released code and trace the tensor $h(\cdot)$ in Eq. (6): if it is the global-average-pooled segment vector broadcast across the $T$ positions instead of per-time-step outputs, the smoothness term is acting on an artifact and the claimed bridge from segment labels to point labels collapses.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that the information gap between noisy segment-level labels and missing point-level labels can be bridged by combining a non-negative PU risk estimator with a time-constraint contrastive loss and a data-centric point selection stage. The PU loss treats known anomalous segments as positives and the rest as unlabeled rather than normal, which prevents the classifier from overfitting to hidden anomalies inside the unlabeled set. The time-constraint loss adds two priors to the per-point anomaly scores: temporal smoothness (adjacent points have similar scores) and separability (positive segments score higher than negative ones), giving the missing point-level supervision a proxy objective. After segment classification, Stage-2 ranks the points of every predicted positive segment and uses a training-free anomaly-rate estimator to pick the threshold, yielding point labels. Across EMG, SMD, PSM, MSL, and SMAP with segment label noise up to 0.6, the paper reports the best pure F1 and PA%K F1 among all compared methods, and its ablation studies attribute the gain to the combination of the PU loss, the TC loss, and the sample selector.

Load-bearing premise

The training losses assume the model produces a real anomaly score for every time point inside a segment, even though the architecture pools the temporal embeddings into one vector before classification.

Editorial extensions

If this is right

  • Point-level detection becomes feasible with event-level annotation: a domain expert only marks that an anomaly occurred within a window, and the model localizes it, cutting labeling cost.
  • Unlabeled data no longer need to be treated as clean normal data; hidden anomalies inside the unlabeled set are handled as PU noise rather than corrupting a normality model.
  • Every component downstream of the temporal embedding—sample selection, PU loss, TC loss, anomaly-rate threshold—is separable, so improvements in any one can be swapped in without redesigning the rest.
  • Lowering the segment-level label noise rate (labeling more true positive segments) tightens the paper's generalization-error upper bound, giving a concrete return on annotation investment.
  • The 11-metric evaluation positions NRdetector as a reference for weak-label time series anomaly detection comparisons, moving beyond pure F1 to PA%K, affiliation, and VUS scores.

Reading between the lines

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

  • The smoothness prior is a bet against point outliers: anomalies that are isolated single-point spikes or abrupt jumps would be penalized by $\mathcal{L}_{\text{smooth}}$, so on datasets dominated by point-wise outliers the method may underperform a model without the TC loss.
  • The static anomaly-rate estimator in Stage-2 assumes the anomaly proportion is roughly constant; in non-stationary or streaming settings, an adaptive rate estimator would be needed to keep the threshold valid.
  • The same 'positive segment plus unlabeled bag plus temporal smoothness' recipe transfers naturally to video anomaly localization or audio event detection, where segment-level positive labels and unlabeled data are the standard annotation format.
  • If the released code shows that $h(\cdot)$ in the TC loss is the pooled vector broadcast to length $T$ rather than true per-time-step outputs, then the point-level supervision is an artifact and the reported gains would need re-attribution to segment-level regularization only.
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

4 major / 5 minor

Summary. The paper considers point-level anomaly detection in multivariate time series when training data has only segment-level labels and a known fraction of positive segments is missing (treated as unlabeled). It proposes NRdetector, a two-stage framework: Stage-1 performs confidence-based sample selection (reliable-negative extraction plus label propagation), then trains a segment classifier with a loss that combines a proposed PU risk estimator and a temporal smoothness/separability term; Stage-2 ranks points inside predicted positive segments and uses an anomaly-rate estimator to set a threshold. Experiments on EMG, SMD, PSM, MSL, and SMAP compare against 13 baselines under several metrics, with ablations of the sample selector, the PU criterion, and the temporal embedding backbone. The central claims are that the proposed PU loss plus time-constraint loss generalizes across noise rates and that the full system outperforms all baselines in point-level F1 and PA%K F1 on all five datasets.

Significance. The problem formulation is practical and timely: real-world anomaly labels are often segment-level and noisy, while deployment requires point-level predictions. The paper ships public code and evaluates on five real-world benchmarks with a broad set of metrics, which is a strength. The proposed two-stage design, combining sample selection, PU learning, and a temporal-constraint loss, is a plausible architecture for this setting. However, the significance is currently limited by two load-bearing gaps: the claimed PU risk estimator in Eq. (5) is not derived and does not match the cited nnPU objective, and the point-level anomaly score h(X) used by the time-constraint loss and Stage-2 is not defined from the architecture in Eq. (2). If these are repaired with a precise derivation and a clear per-point scoring definition, the empirical comparison could be valuable; as written, the central mechanism is not reproducible from the text alone.

major comments (4)
  1. [Section 4.2.3, Eq. (5)] Equation (5) is presented as the Non-negative PU Risk Estimator and is justified by an upper-bound result from [62], but it does not match the nnPU objective: the nnPU risk has the form pi_P E_P[ell(f(X),1)] + max(0, E_U[ell(f(X),0)] - pi_P E_P[ell(f(X),0)]) for a surrogate loss ell, whereas Eq. (5) averages the sigmoid output f(X) itself and takes absolute deviations from 1 and pi_P. No derivation links the two, and the cited proposition from [62] is not stated. Because R_pu is the main segment-level objective in Eq. (9), the theoretical basis of Stage-1 is currently asserted rather than established.
  2. [Section 4.2.3, Eq. (6), and Section 4.3] The point-level anomaly score h(X) used by Eq. (6) and by Stage-2 is never defined. Equation (2) builds a segment representation X = GAP(h_1,...,h_T) in R^d and produces one scalar S = sigma(w^T X); no equation specifies a T-dimensional scalar score per segment. If h(X_i)_j in Eq. (6) is the j-th d-dimensional temporal embedding, then the squared difference is not a scalar loss; if it is a scalar point score, the text does not say how it is computed from the pooled representation. Since L_smooth and the Stage-2 ranking both require per-point scores, the central point-level mechanism has no well-defined object as written. Please specify h(X) precisely, for example as a per-timestep output head applied before pooling, and correct Eq. (6) accordingly.
  3. [Section 5.2, Table 10, and Section 4.3] The claim that Stage-2 automatically determines the threshold is only partially supported: Table 10 fixes anomaly_ratio per dataset, and Section 4.3 describes k as a user-specified hyperparameter. Similarly, the class prior pi_P in Eq. (5) is set per dataset in Table 10. No validation or selection protocol for these per-dataset values is described, so the pipeline is not fully self-contained and the 'training-free automated estimator' does not remove the dependence on the fitted k. At minimum, the paper should report sensitivity to k and pi_P on all datasets or justify fixed values, and should clarify what the HOC estimator actually estimates given k.
  4. [Tables 2-4 and Section 5.3] All experimental tables report single point estimates with no standard deviations, confidence intervals, or number of seeds. This matters for the headline claim: on SMD in Table 2, NRdetector's F1 (0.1092) is close to WETAS (0.1020) and TreeMIL (0.0999), and in Table 4 under noise rate 0.0 the ranking is not consistently first. Without repeated-run statistics, the statement that NRdetector achieves the best results under the pure F1 score on all benchmark datasets is not statistically grounded.
minor comments (5)
  1. [Section 4.2.1] The symbol S in Eq. (2) is a segment-level score, yet Section 4.3 refers to 'anomaly scores S corresponding to each point'; the notation should be separated to avoid ambiguity.
  2. [Section 5.3] The text says 'NRdetector still outperforms WEATS and TreeMIL' where the intended baseline name is WETAS; please correct this typo.
  3. [Table 3 caption] The caption says 'Aff-P and Aff-P are the precision/recall pair' but the column names are Aff-P and Aff-R; the caption should be corrected.
  4. [Table 10] The hyperparameter is listed as 'anoamly_ratio' in the table; this should be 'anomaly_ratio'.
  5. [Section 4.2.2, Eq. (3)] The quantities Omega, Phi, and mu in Eq. (3) are not defined in the main text; the Katz index is also cited but not defined, so the sample selector is hard to reproduce from the paper alone.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the headline benchmark comparisons are self-contained and the internal loss terms do not reduce to the evaluation targets by construction.

full rationale

The central claim of the paper is an empirical comparison: NRdetector outperforms 13 baselines on five real-world datasets under point-level F1 and PA%K metrics. That comparison is self-contained against external benchmarks and does not reuse the paper's own fitted values as ground truth. The PU loss in Eq. (5) and the time-constraint loss in Eqs. (6)-(8) are regularizers defined on segment scores and on an asserted T-dimensional point-score h(X); minimizing them does not by construction maximize point-level F1 on held-out data, so no reported prediction is statistically forced by the loss definitions. The Stage-2 threshold is set by the HOC estimator [68], which is a self-citation, but it is used as an external peer-reviewed tool with its own assumptions, not as a uniqueness theorem and not as a definition of the target labels; the paper does not reduce the point-level prediction to HOC's output by fiat. A real specification gap exists in Section 4.2.3: the text calls the previous layer 'the T-dimension output h(X)' and Eq. (6) uses h(X_i)_j as a point score, while Eq. (2) pools the temporal embeddings into X before classification, leaving the T point scores undefined. This is a missing definition and a reproducibility concern, not a circular reduction: no equation equates the point-level F1 result to the loss inputs by construction. Per-dataset hyperparameters such as prior and anomaly_ratio in Table 10 are conventional tuning choices, and they are not relabeled as predictions. The self-citations [67]-[69] support auxiliary bounds and tools rather than the central derivation, and they are not invoked to forbid alternative methods. Therefore the derivation chain is not circular.

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

The central claim depends on several fitted quantities: per-dataset class priors, per-dataset initial anomaly ratios, and tuned loss weights. The PU loss in Eq. (5) is asserted but not derived, and the point-level score definition is ambiguous. No new entities are introduced. The paper's contribution is an engineering combination of known building blocks rather than a derivation, so the ledger is dominated by tuning choices and domain assumptions about temporal smoothness.

free parameters (5)
  • Segment class prior pi_P = EMG 0.25, SMD 0.8, PSM 0.4, MSL 0.5, SMAP 0.5 (Table 10)
    Used in the PU loss Eq. (5); no estimation method is given, values appear chosen per dataset.
  • Initial anomaly ratio k (anomaly_ratio) = EMG 0.65, SMD 0.15, PSM 0.6, MSL 0.8, SMAP 0.9 (Table 10)
    Used to generate pseudo point labels before HOC estimation in Stage-2; despite the automated estimator, the initial ratio is set per dataset.
  • TC loss weights lambda1, lambda2 = 8e-5 each (Section 5.2)
    Described as chosen for best performance; affects the balance between smoothness and separability.
  • Sample selector parameters m, lambda0 = m=4, lambda0=0.32 (Appendix C.2)
    Control how many unlabeled segments are filtered as reliable negatives; hand-set.
  • Temporal embedding dimension d_model = 64 (Appendix C.2)
    Architecture choice; moderate impact on learned representations.
assumptions (5)
  • domain assumption Positive segments are labeled uniformly at random and independently of features; unlabeled data are drawn from the true marginal distribution.
    Standard PU assumption stated in Section 3.3, needed for the PU risk and sample selector; violations change the noise model.
  • domain assumption Point-level anomaly scores vary smoothly within normal and abnormal segments, and abnormal segments have higher average scores than normal segments.
    Motivated by Figure 1 and used in Section 4.2.3 to design the TC loss; it can fail for isolated spike anomalies.
  • domain assumption The noise-rate estimator HOC can recover the true anomaly rate from pseudo point labels and features.
    Invoked in Section 4.3 via HOC [68]; requires anchor points or clusterability in feature space, not verified on these datasets.
  • domain assumption There are no false positive segment labels, i.e., e0 = P(tilde Y=1 | Y=0) = 0.
    The PU formulation in Section 3.3 assumes labeled positives are true positives; verification errors would violate the model.
  • standard math Each feature X belongs to one true class with probability 1.
    Used in Appendix B.1 to derive the generalization bound; standard in noisy-label analysis but an assumption about the data.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Noise-Resilient Point-wise Anomaly Detection in Time Series Using Weak Segment Labels." pith.science (2026). https://pith.science/paper/LTMK3ZBW

@misc{pith2026250111959,
  author       = {Pith},
  title        = {Pith review of: Noise-Resilient Point-wise Anomaly Detection in Time Series Using Weak Segment Labels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LTMK3ZBW}},
  note         = {Machine review of arXiv:2501.11959}
}
read the original abstract

Detecting anomalies in temporal data has gained significant attention across various real-world applications, aiming to identify unusual events and mitigate potential hazards. In practice, situations often involve a mix of segment-level labels (detected abnormal events with segments of time points) and unlabeled data (undetected events), while the ideal algorithmic outcome should be point-level predictions. Therefore, the huge label information gap between training data and targets makes the task challenging. In this study, we formulate the above imperfect information as noisy labels and propose NRdetector, a noise-resilient framework that incorporates confidence-based sample selection, robust segment-level learning, and data-centric point-level detection for multivariate time series anomaly detection. Particularly, to bridge the information gap between noisy segment-level labels and missing point-level labels, we develop a novel loss function that can effectively mitigate the label noise and consider the temporal features. It encourages the smoothness of consecutive points and the separability of points from segments with different labels. Extensive experiments on real-world multivariate time series datasets with 11 different evaluation metrics demonstrate that NRdetector consistently achieves robust results across multiple real-world datasets, outperforming various baselines adapted to operate in our setting.

Figures

Figures reproduced from arXiv: 2501.11959 by the authors.

Figure 1
Figure 1. Illustration of the insights. The x-axis in (a) and [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The workflow of the NRdetector framework. NRdetector consists of two main stages: coarse-grained PU learning and [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Parameter sensitivity studies of hyper-parameters, [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

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. Cluster-Aware Causal Mixer for Online Anomaly Detection in Multivariate Time Series

    cs.LG 2025-05 conditional novelty 6.0 of 10

    A cluster-aware causal MLP-mixer with sequential evidence-accumulation scoring achieves state-of-the-art best-F1 on six multivariate time-series anomaly detection benchmarks.

Reference graph

Works this paper leans on

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

  1. [62]

    Jiaqi Zhang, Zhenzhen Wang, Junsong Yuan, and Yap-Peng Tan. 2017. Positive and unlabeled learning for anomaly detection with multi-features. In Proceedings of the 25th ACM international conference on Multimedia . 854–862

    work page 2017

  2. [1]

    Ahmed Abdulaal, Zhuanghua Liu, and Tomer Lancewicki. 2021. Practical ap- proach to asynchronous multivariate time series anomaly detection and localiza- tion. In Proceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining. 2485–2494

    2021

  3. [2]

    Peter L Bartlett, Michael I Jordan, and Jon D McAuliffe. 2006. Convexity, classifi- cation, and risk bounds. J. Amer. Statist. Assoc. 101, 473 (2006), 138–156

    work page 2006

  4. [3]

    Jessa Bekker and Jesse Davis. 2020. Learning from positive and unlabeled data: A survey. Machine Learning 109 (2020), 719–760

    work page 2020

  5. [4]

    Chris U Carmona, François-Xavier Aubet, Valentin Flunkert, and Jan Gasthaus

    work page

  6. [5]

    Feiyi Chen, Zhen Qing, Yingying Zhang, Shuiguang Deng, Yi Xiao, Guansong Pang, and Qingsong Wen. 2023. LARA: A Light and Anti-overfitting Retraining Approach for Unsupervised Anomaly Detection. arXiv preprint arXiv:2310.05668 (2023)

    work page Pith review arXiv 2023

  7. [6]

    Xuxi Chen, Wuyang Chen, Tianlong Chen, Ye Yuan, Chen Gong, Kewei Chen, and Zhangyang Wang. 2020. Self-pu: Self boosted and calibrated positive-unlabeled training. In International Conference on Machine Learning . PMLR, 1510–1519

    work page 2020

  8. [7]

    Zahra Zamanzadeh Darban, Geoffrey I Webb, Shirui Pan, Charu C Aggarwal, and Mahsa Salehi. 2025. CARLA: Self-supervised contrastive representation learning for time series anomaly detection. Pattern Recognition 157 (2025), 110874

    work page 2025

Show all 75 references
  1. [8]

    Mariana Caravanti de Souza, Bruno Magalhães Nogueira, Rafael Geraldeli Rossi, Ricardo Marcondes Marcacini, Brucce Neves Dos Santos, and Solange Oliveira Rezende. 2022. A network-based positive and unlabeled learning approach for fake news detection. Machine learning 111, 10 (2...

    2022

  2. [9]

    Marthinus C Du Plessis, Gang Niu, and Masashi Sugiyama. 2014. Analysis of learning from positive and unlabeled data. Advances in neural information processing systems 27 (2014)

    2014

  3. [10]

    Eman Abd Elaziz, Radwa Fathalla, and Mohamed Shaheen. 2023. Deep reinforce- ment learning for data-efficient weakly supervised business process anomaly detection. Journal of Big Data 10, 1 (2023), 33

    2023

  4. [11]

    Jingkun Gao, Xiaomin Song, Qingsong Wen, Pichao Wang, Liang Sun, and Huan Xu. 2020. RobustTAD: Robust time series anomaly detection via decomposition and convolutional neural networks. arXiv preprint arXiv:2002.09545 (2020)

    2020 arXiv

  5. [12]

    Bo Han, Quanming Yao, Xingrui Yu, Gang Niu, Miao Xu, Weihua Hu, Ivor Tsang, and Masashi Sugiyama. 2018. Co-teaching: Robust training of deep neural net- works with extremely noisy labels. Advances in neural information processing systems 31 (2018)

    2018

  6. [13]

    Wenpeng Hu, Ran Le, Bing Liu, Feng Ji, Jinwen Ma, Dongyan Zhao, and Rui Yan. 2021. Predictive adversarial learning from positive and unlabeled data. In Proceedings of the AAAI conference on artificial intelligence , Vol. 35. 7806–7814

    2021

  7. [14]

    Alexis Huet, Jose Manuel Navarro, and Dario Rossi. 2022. Local evaluation of time series anomaly detection algorithms. InProceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 635–645

    2022

  8. [15]

    Kyle Hundman, Valentino Constantinou, Christopher Laporte, Ian Colwell, and Tom Soderstrom. 2018. Detecting spacecraft anomalies using lstms and nonpara- metric dynamic thresholding. In Proceedings of the 24th ACM SIGKDD interna- tional conference on knowledge discovery & data...

    2018

  9. [16]

    Dongmin Kim, Sunghyun Park, and Jaegul Choo. 2023. When Model Meets New Normals: Test-time Adaptation for Unsupervised Time-series Anomaly Detection. arXiv preprint arXiv:2312.11976 (2023)

    2023 arXiv

  10. [17]

    Siwon Kim, Kukjin Choi, Hyun-Soo Choi, Byunghan Lee, and Sungroh Yoon. 2022. Towards a rigorous evaluation of time-series anomaly detection. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 36. 7194–7201

    2022

  11. [18]

    Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  12. [19]

    Ryuichi Kiryo, Gang Niu, Marthinus C Du Plessis, and Masashi Sugiyama. 2017. Positive-unlabeled learning with non-negative risk estimator. Advances in neural information processing systems 30 (2017)

    2017

  13. [20]

    Chih-Yu Andrew Lai, Fan-Keng Sun, Zhengqi Gao, Jeffrey H Lang, and Duane Boning. 2024. Nominality score conditioned time series anomaly detection by point/sequential reconstruction. Advances in Neural Information Processing Systems 36 (2024)

    2024

  14. [21]

    Kwei-Herng Lai, Daochen Zha, Junjie Xu, Yue Zhao, Guanchu Wang, and Xia Hu. 2021. Revisiting time series outlier detection: Definitions and benchmarks. In Thirty-fifth conference on neural information processing systems datasets and benchmarks track (round 1)

    2021

  15. [22]

    Dongha Lee, Sehun Yu, Hyunjun Ju, and Hwanjo Yu. 2021. Weakly supervised temporal anomaly segmentation with dynamic time warping. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 7355–7364

    2021

  16. [23]

    Longyuan Li, Junchi Yan, Qingsong Wen, Yaohui Jin, and Xiaokang Yang. 2022. Learning robust deep state space for unsupervised anomaly detection in con- taminated time-series. IEEE Transactions on Knowledge and Data Engineering (2022)

    2022

  17. [24]

    Xiaoli Li and Bing Liu. 2003. Learning to classify texts using positive and unlabeled data. In IJCAI, Vol. 3. 587–592

    2003

  18. [25]

    Min Lin, Qiang Chen, and Shuicheng Yan. 2013. Network in network. arXiv preprint arXiv:1312.4400 (2013)

    2013 arXiv

  19. [26]

    Bing Liu, Wee Sun Lee, Philip S Yu, and Xiaoli Li. 2002. Partially supervised classification of text documents. In ICML, Vol. 2. Sydney, NSW, 387–394

    2002

  20. [27]

    Chen Liu, Shibo He, Haoyu Liu, and Shizhong Li. 2024. TreeMIL: A Multi- instance Learning Framework for Time Series Anomaly Detection with Inexact Supervision. arXiv preprint arXiv:2401.11235 (2024)

    2024 arXiv

  21. [28]

    Tongliang Liu and Dacheng Tao. 2015. Classification with noisy labels by impor- tance reweighting. IEEE Transactions on pattern analysis and machine intelligence 38, 3 (2015), 447–461

    2015

  22. [29]

    Yang Liu and Hongyi Guo. 2020. Peer loss functions: Learning from noisy labels without knowing noise rates. In International conference on machine learning . PMLR, 6226–6236

    2020

  23. [30]

    Sergey Lobov, Nadia Krilova, Innokentiy Kastalskiy, Victor Kazantsev, and Va- leri A Makarov. 2018. Latent factors limiting the performance of sEMG-interfaces. Sensors 18, 4 (2018), 1122

    2018

  24. [31]

    Chuan Luo, Pu Zhao, Chen Chen, Bo Qiao, Chao Du, Hongyu Zhang, Wei Wu, Shaowei Cai, Bing He, Saravanakumar Rajmohan, et al. 2021. Pulns: Positive- unlabeled learning with effective negative sample selector. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol...

    2021

  25. [32]

    Shuangxun Ma and Ruisheng Zhang. 2017. PU-LP: A novel approach for positive and unlabeled learning by label propagation. In2017 IEEE International Conference on Multimedia & Expo Workshops (ICMEW) . IEEE, 537–542

    2017

  26. [33]

    Vinod Nair and Geoffrey E Hinton. 2010. Rectified linear units improve re- stricted boltzmann machines. In Proceedings of the 27th international conference on machine learning (ICML-10) . 807–814

    2010

  27. [34]

    Minh Nhut Nguyen, Xiao-Li Li, and See-Kiong Ng. 2011. Positive unlabeled learn- ing for time series classification. In Twenty-Second International Joint Conference on Artificial Intelligence. Citeseer

    2011

  28. [35]

    Aaron van den Oord, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew Senior, and Koray Kavukcuoglu

  29. [36]

    John Paparrizos, Paul Boniol, Themis Palpanas, Ruey S Tsay, Aaron Elmore, and Michael J Franklin. 2022. Volume under the surface: a new accuracy evaluation measure for time-series anomaly detection. Proceedings of the VLDB Endowment 15, 11 (2022), 2774–2787

    2022

  30. [37]

    Daehyung Park, Yuuna Hoshi, and Charles C Kemp. 2018. A multimodal anomaly detector for robot-assisted feeding using an lstm-based variational autoencoder. IEEE Robotics and Automation Letters 3, 3 (2018), 1544–1551

    2018

  31. [38]

    Giorgio Patrini, Alessandro Rozza, Aditya Krishna Menon, Richard Nock, and Lizhen Qu. 2017. Making deep neural networks robust to label noise: A loss correction approach. In Proceedings of the IEEE conference on computer vision and pattern recognition. 1944–1952

    2017

  32. [39]

    Lorenzo Perini, Vincent Vercruyssen, and Jesse Davis. 2023. Learning from Positive and Unlabeled Multi-Instance Bags in Anomaly Detection. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 1897–1906

    2023

  33. [40]

    Lukas Ruff, Robert Vandermeulen, Nico Goernitz, Lucas Deecke, Shoaib Ahmed Siddiqui, Alexander Binder, Emmanuel Müller, and Marius Kloft. 2018. Deep one-class classification. In International conference on machine learning . PMLR, 4393–4402

    2018

  34. [41]

    Artem Ryzhikov, Maxim Borisyak, Andrey Ustyuzhanin, and Denis Derkach

  35. [42]

    Lifeng Shen, Zhuocong Li, and James Kwok. 2020. Timeseries anomaly detection using temporal hierarchical one-class network. Advances in Neural Information Processing Systems 33 (2020), 13016–13026

    2020

  36. [43]

    Ya Su, Youjian Zhao, Chenhao Niu, Rong Liu, Wei Sun, and Dan Pei. 2019. Robust anomaly detection for multivariate time series through stochastic recurrent neural network. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining . 2828–2837

    2019

  37. [44]

    PeerJ Computer Science 7 (2021), e757

    NFAD: fixing anomaly detection using normalizing flows. PeerJ Computer Science 7 (2021), e757

    2021

  38. [45]

    Yuting Sun, Guansong Pang, Guanhua Ye, Tong Chen, Xia Hu, and Hongzhi Yin. 2023. Unraveling theAnomaly’in Time Series Anomaly Detection: A Self- supervised Tri-domain Solution. arXiv preprint arXiv:2311.11235 (2023)

    2023 arXiv

  39. [46]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems 30 (2017)

    2017

  40. [47]

    Waqas Sultani, Chen Chen, and Mubarak Shah. 2018. Real-world anomaly de- tection in surveillance videos. In Proceedings of the IEEE conference on computer vision and pattern recognition . 6479–6488

    2018

  41. [48]

    Jiaheng Wei, Hangyu Liu, Tongliang Liu, Gang Niu, Masashi Sugiyama, and Yang Liu. 2021. To smooth or not? when label smoothing meets noisy labels. arXiv preprint arXiv:2106.04149 (2021). Noise-Resilient Point-wise Anomaly Detection in Time Series Using Weak Segment Labels KDD ...

    2021 arXiv

  42. [49]

    Jiaheng Wei, Zhaowei Zhu, Hao Cheng, Tongliang Liu, Gang Niu, and Yang Liu

  43. [50]

    Rui Wang, Xudong Mou, Renyu Yang, Kai Gao, Pin Liu, Chongwei Liu, Tianyu Wo, and Xudong Liu. 2024. CutAddPaste: Time Series Anomaly Detection by Exploit- ing Abnormal Knowledge. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 3176–3187

    2024

  44. [51]

    Haixu Wu, Jiehui Xu, Jianmin Wang, and Mingsheng Long. 2021. Autoformer: De- composition transformers with auto-correlation for long-term series forecasting. Advances in Neural Information Processing Systems 34 (2021), 22419–22430

    2021

  45. [52]

    Haowen Xu, Wenxiao Chen, Nengwen Zhao, Zeyan Li, Jiahao Bu, Zhihan Li, Ying Liu, Youjian Zhao, Dan Pei, Yang Feng, et al. 2018. Unsupervised anomaly detection via variational auto-encoder for seasonal kpis in web applications. In Proceedings of the 2018 world wide web conferen...

    2018

  46. [53]

    arXiv preprint arXiv:2110.12088 (2021)

    Learning with noisy labels revisited: A study using real-world human annotations. arXiv preprint arXiv:2110.12088 (2021)

    2021 arXiv

  47. [54]

    Haixu Wu, Tengge Hu, Yong Liu, Hang Zhou, Jianmin Wang, and Mingsheng Long. 2022. Timesnet: Temporal 2d-variation modeling for general time series analysis. arXiv preprint arXiv:2210.02186 (2022)

    2022 arXiv

  48. [55]

    Yiyuan Yang, Chaoli Zhang, Tian Zhou, Qingsong Wen, and Liang Sun. 2023. DCdetector: Dual Attention Contrastive Representation Learning for Time Se- ries Anomaly Detection. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

    2023

  49. [56]

    Chen Zhang, Guorong Li, Yuankai Qi, Shuhui Wang, Laiyun Qing, Qingming Huang, and Ming-Hsuan Yang. 2023. Exploiting Completeness and Uncertainty of Pseudo Labels for Weakly Supervised Video Anomaly Detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Patt...

    2023

  50. [57]

    Hongzuo Xu, Yijie Wang, Songlei Jian, Qing Liao, Yongjun Wang, and Guansong Pang. 2024. Calibrated one-class classification for unsupervised time series anomaly detection. IEEE Transactions on Knowledge and Data Engineering (2024)

    2024

  51. [58]

    Jiehui Xu, Haixu Wu, Jianmin Wang, and Mingsheng Long. 2021. Anomaly transformer: Time series anomaly detection with association discrepancy. arXiv preprint arXiv:2110.02642 (2021)

    2021 arXiv

  52. [59]

    Kexin Zhang, Qingsong Wen, Chaoli Zhang, Liang Sun, and Yong Liu. 2022. Time Series Anomaly Detection using Skip-Step Contrastive Predictive Coding. In NeurIPS 2022 Workshop: Self-Supervised Learning-Theory and Practice

    2022

  53. [60]

    Shenglin Zhang, Chenyu Zhao, Yicheng Sui, Ya Su, Yongqian Sun, Yuzhi Zhang, Dan Pei, and Yizhe Wang. 2021. Robust KPI anomaly detection for large-scale software services with partial labels. In 2021 IEEE 32nd International Symposium on Software Reliability Engineering (ISSRE) ...

    2021

  54. [61]

    Chaoli Zhang, Tian Zhou, Qingsong Wen, and Liang Sun. 2022. TFAD: A de- composition time series anomaly detection architecture with time-frequency analysis. In Proceedings of the 31st ACM International Conference on Information & Knowledge Management. 2497–2507

    2022

  55. [63]

    Tian Zhou, Ziqing Ma, Qingsong Wen, Xue Wang, Liang Sun, and Rong Jin

  56. [64]

    Tian Zhou, Peisong Niu, Xue Wang, Liang Sun, and Rong Jin. 2023. One Fits All: Power General Time Series Analysis by Pretrained LM. arXiv preprint arXiv:2302.11939 (2023)

    2023 arXiv

  57. [65]

    Yingying Zhang, Zhengxiong Guan, Huajie Qian, Leili Xu, Hengbo Liu, Qingsong Wen, Liang Sun, Junwei Jiang, Lunting Fan, and Min Ke. 2021. CloudRCA: A root cause analysis framework for cloud computing platforms. In Proceedings of the 30th ACM International Conference on Informa...

    2021

  58. [66]

    Yunrui Zhao, Qianqian Xu, Yangbangyan Jiang, Peisong Wen, and Qingming Huang. 2022. Dist-pu: Positive-unlabeled learning from a label distribution perspective. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 14461–14470

    2022

  59. [67]

    Zhaowei Zhu, Tianyi Luo, and Yang Liu. 2021. The rich get richer: Disparate impact of semi-supervised learning. arXiv preprint arXiv:2110.06282 (2021)

    2021 arXiv

  60. [68]

    Zhaowei Zhu, Yiwen Song, and Yang Liu. 2021. Clusterability as an alternative to anchor points when learning with noisy labels. In International Conference on Machine Learning. PMLR, 12912–12923

    2021

  61. [69]

    Zhaowei Zhu, Jialu Wang, Hao Cheng, and Yang Liu. 2023. Unmasking and improving data credibility: A study with datasets for training harmless language models. arXiv preprint arXiv:2311.11202 (2023)

    2023 arXiv

  62. [70]

    Zhi-Hua Zhou. 2004. Multi-instance learning: A survey. Department of Computer Science & Technology, Nanjing University, Tech. Rep 1 (2004)

    2004

  63. [71]

    Zhaowei Zhu, Tongliang Liu, and Yang Liu. 2021. A second-order approach to learning with instance-dependent label noise. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 10113–10123

    2021

  64. [75]

    Zhangchi Zhu, Lu Wang, Pu Zhao, Chao Du, Wei Zhang, Hang Dong, Bo Qiao, Qingwei Lin, Saravan Rajmohan, and Dongmei Zhang. 2023. Robust Positive- Unlabeled Learning via Noise Negative Sample Self-correction. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discover...

    2023

  65. [2016]

    Wavenet: A generative model for raw audio.arXiv preprint arXiv:1609.03499 (2016)

    2016 arXiv

  66. [2021]

    arXiv preprint arXiv:2107.07702 (2021)

    Neural contextual anomaly detection for time series. arXiv preprint arXiv:2107.07702 (2021)

    2021 arXiv

  67. [2022]

    In International Conference on Machine Learning

    Fedformer: Frequency enhanced decomposed transformer for long-term series forecasting. In International Conference on Machine Learning. PMLR, 27268– 27286

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.