Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
97 tokens/sec
GPT-4o
53 tokens/sec
Gemini 2.5 Pro Pro
44 tokens/sec
o3 Pro
5 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Towards a General Time Series Anomaly Detector with Adaptive Bottlenecks and Dual Adversarial Decoders (2405.15273v3)

Published 24 May 2024 in cs.LG

Abstract: Time series anomaly detection plays a vital role in a wide range of applications. Existing methods require training one specific model for each dataset, which exhibits limited generalization capability across different target datasets, hindering anomaly detection performance in various scenarios with scarce training data. Aiming at this problem, we propose constructing a general time series anomaly detection model, which is pre-trained on extensive multi-domain datasets and can subsequently apply to a multitude of downstream scenarios. The significant divergence of time series data across different domains presents two primary challenges in building such a general model: (1) meeting the diverse requirements of appropriate information bottlenecks tailored to different datasets in one unified model, and (2) enabling distinguishment between multiple normal and abnormal patterns, both are crucial for effective anomaly detection in various target scenarios. To tackle these two challenges, we propose a General time series anomaly Detector with Adaptive Bottlenecks and Dual Adversarial Decoders (DADA), which enables flexible selection of bottlenecks based on different data and explicitly enhances clear differentiation between normal and abnormal series. We conduct extensive experiments on nine target datasets from different domains. After pre-training on multi-domain data, DADA, serving as a zero-shot anomaly detector for these datasets, still achieves competitive or even superior results compared to those models tailored to each specific dataset.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (70)
  1. Anomaly detection in finance: Editors’ introduction. In ADF@KDD, 2017.
  2. Anomaly detection for iot time-series data: A survey. IEEE Internet Things J., 2020.
  3. Anomaly detection in time series with robust variational quasi-recurrent autoencoders. In ICDE, 2022.
  4. Pipeline safety early warning by multifeature-fusion CNN and lightgbm analysis of signals from distributed optical fiber sensors. IEEE Trans. Instrum. Meas., 2021.
  5. DISTDET: A cost-effective distributed cyber threat detection system. In USENIX Security Symposium, 2023a.
  6. Wrongdoing monitor: A graph-based behavioral anomaly detection in cyber security. IEEE Trans. Inf. Forensics Secur., 2022.
  7. A study on the impact of memory dos attacks on cloud applications and exploring real-time detection schemes. IEEE/ACM Trans. Netw., 2022.
  8. ECGGAN: A framework for effective and interpretable electrocardiogram anomaly detection. In KDD, 2023a.
  9. Markov models for anomaly detection in wireless body area networks for secure health monitoring. IEEE J. Sel. Areas Commun., 2021.
  10. Anomaly transformer: Time series anomaly detection with association discrepancy. In ICLR, 2022.
  11. Robust anomaly detection for multivariate time series through stochastic recurrent neural network. In KDD, 2019a.
  12. ModernTCN: A modern pure convolution structure for general time series analysis. In ICLR, 2024.
  13. Drift doesn’t matter: Dynamic decomposition with diffusion reconstruction for unstable multivariate time series anomaly detection. In NeurIPS, 2023b.
  14. Unsupervised time series outlier detection with diversity-driven convolutional ensembles - extended version. CoRR, 2021.
  15. Dcdetector: Dual attention contrastive representation learning for time series anomaly detection. In KDD, 2023.
  16. Self-supervised learning for time series analysis: Taxonomy, progress, and prospects. CoRR, 2023.
  17. Deep learning for anomaly detection: A survey. CoRR, 2019.
  18. How does information bottleneck help deep learning? In ICML, 2023.
  19. Unitime: A language-empowered unified model for cross-domain time series forecasting. In WWW, 2024a.
  20. A comparative study on unsupervised anomaly detection for time series: Experiments and analysis. CoRR, 2022.
  21. LOF: identifying density-based local outliers. In SIGMOD Conference, 2000.
  22. Loop: local outlier probabilities. In CIKM, 2009.
  23. A robust outlier detection scheme for large data sets. Proc. Pacific-Asia Conf. Knowledge Discovery and Data Mining, 01 2002a.
  24. Enhancing effectiveness of outlier detections for low density patterns. In PAKDD, 2002b.
  25. Matrix profile I: all pairs similarity joins for time series: A unifying view that includes motifs, discords and shapelets. In ICDM, 2016.
  26. Isolation forest. In ICDM, 2008.
  27. Support vector method for novelty detection. In NIPS, 1999.
  28. Imagenet classification with deep convolutional neural networks. In NIPS, 2012.
  29. A multimodal anomaly detector for robot-assisted feeding using an lstm-based variational autoencoder. IEEE Robotics Autom. Lett., 2018.
  30. Multivariate time series anomaly detection and interpretation using hierarchical inter-metric and temporal embedding. In KDD, 2021a.
  31. Beatgan: Anomalous rhythm detection using adversarially generated time series. In IJCAI, 2019.
  32. MAD-GAN: multivariate anomaly detection for time series data with generative adversarial networks. In ICANN (4), 2019.
  33. f-anogan: Fast unsupervised anomaly detection with generative adversarial networks. Medical Image Anal., 2019.
  34. Learning graph structures with transformer for multivariate time-series anomaly detection in iot. IEEE Internet Things J., 2022.
  35. Deep learning for anomaly detection: A review. ACM Comput. Surv., 2022.
  36. A time series is worth 64 words: Long-term forecasting with transformers. In ICLR, 2023.
  37. Simmtm: A simple pre-training framework for masked time-series modeling. In NeurIPS, 2023b.
  38. Time series contrastive learning with information-aware augmentations. In AAAI, 2023.
  39. Ts2vec: Towards universal representation of time series. In AAAI, 2022.
  40. Large language models are zero-shot time series forecasters. In NeurIPS, 2023.
  41. One fits all: Power general time series analysis by pretrained LM. In NeurIPS, 2023.
  42. Timegpt-1. CoRR, 2023.
  43. Unified training of universal time series forecasting transformers. CoRR, 2024.
  44. Forecastpfn: Synthetically-trained zero-shot forecasting. In NeurIPS, 2023.
  45. Timer: Transformers for time series analysis at scale. CoRR, 2024b.
  46. MOMENT: A family of open time-series foundation models. In ICML, 2024.
  47. Anomaly detection using autoencoders with nonlinear dimensionality reduction. In MLSDA@PRICAI, 2014.
  48. Unsupervised domain adaptation by backpropagation. In ICML, 2015.
  49. Anomaly detection in streams with extreme value theory. In KDD, 2017.
  50. Monash time series forecasting archive. In NeurIPS Datasets and Benchmarks, 2021.
  51. Multivariate time series anomaly detection and interpretation using hierarchical inter-metric and temporal embedding. In KDD, 2021b.
  52. Exathlon: A benchmark for explainable anomaly detection over time series. Proc. VLDB Endow., 2021.
  53. Time-series anomaly detection service at microsoft. In KDD, 2019.
  54. Collecting complex activity datasets in highly rich networked sensor environments. In INSS, 2010.
  55. A comparative study of HTM and other neural network models for online sequence learning with streaming data. In IJCNN, 2016.
  56. Markusthill/mgab: The mackey-glass anomaly benchmark (version v1.0.1). https://doi.org/10.5281/zenodo.3762385, April 2020.
  57. The impact of the mit-bih arrhythmia database. IEEE Engineering in Medicine and Biology Magazine, 2001.
  58. Improved detection and classification of arrhythmias in noise-corrupted electrocardiograms using contextual information. [1990] Proceedings Computers in Cardiology, 1990.
  59. S5-a labeled anomaly detection dataset, version 1.0 (16m), 2015.
  60. Robust anomaly detection for multivariate time series through stochastic recurrent neural network. In KDD, 2019b.
  61. Detecting spacecraft anomalies using lstms and nonparametric dynamic thresholding. In KDD, 2018.
  62. Swat: a water treatment testbed for research and training on ICS security. In CySWATER@CPSWeek, 2016.
  63. Practical approach to asynchronous multivariate time series anomaly detection and localization. In KDD, 2021.
  64. Revisiting time series outlier detection: Definitions and benchmarks. In NeurIPS Datasets and Benchmarks, 2021.
  65. A novel anomaly detection scheme based on principal component classifier. In Proceedings of International Conference on Data Mining, 2003.
  66. Histogram-based outlier score (hbos): A fast unsupervised anomaly detection algorithm. KI-2012: poster and demo track, 2012.
  67. Tomás Pevný. Loda: Lightweight on-line detector of anomalies. Mach. Learn., 2016.
  68. Deep autoencoding gaussian mixture model for unsupervised anomaly detection. In ICLR, 2018.
  69. Local evaluation of time series anomaly detection algorithms. In KDD, 2022.
  70. Adaptive mixtures of local experts. Neural Comput., 1991.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (8)
  1. Qichao Shentu (1 paper)
  2. Beibu Li (1 paper)
  3. Kai Zhao (160 papers)
  4. Zhongwen Rao (11 papers)
  5. Lujia Pan (27 papers)
  6. Bin Yang (320 papers)
  7. Chenjuan Guo (48 papers)
  8. Yang Shu (17 papers)

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets