Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

AnomalyTCN: Efficient contrastive-based time series anomaly detection with pure convolution structure.

Donghao Luo1, Xue Wang1

  • 1Department of Precision Instrument, Tsinghua University, Beijing, 100084, China.

Neural Networks : the Official Journal of the International Neural Network Society
|March 25, 2026
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Difference from Background: Limit of Detection01:05

Difference from Background: Limit of Detection

8.9K
The limit of detection (LOD) is the smallest amount of analyte that can be distinguished from the background noise. The LOD value corresponds to the concentration at which the analyte signal is three times larger than the standard deviation of the blank signal. Below this value, the analyte signal cannot be differentiated from the background noise. It is calculated by dividing the calibration slope by 3 times the standard deviation of the blank signals.
The LOD indicates the presence or absence...
8.9K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

OBIMD: A Multi-modal Dataset for Contextual Interpretation of Oracle Bone Inscriptions.

Scientific data·2026
Same author

Bench-Stable Carbamoylzinc Pivalates for Modular Access to Amides, Ureas, and Thiocarbamates.

Journal of the American Chemical Society·2025
Same author

Dynamic Frame Interpolation in Wavelet Domain.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2023
Same author

Maternal inheritance of glucose intolerance via oocyte TET3 insufficiency.

Nature·2022
Same author

Direct Access to Allenylphosphine Oxides via a Metal Free Coupling of Propargylic Substrates with P(O)H Compounds.

Organic letters·2019
Same author

Therapeutic effect of methyl salicylate 2-O-β-d-lactoside on LPS-induced acute lung injury by inhibiting TAK1/NF-kappaB phosphorylation and NLRP3 expression.

International immunopharmacology·2016

We introduce AnomalyTCN, an efficient, attention-free method for contrastive-based time series anomaly detection. This novel approach maintains high performance while significantly reducing computational costs and memory usage.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Data Science

Background:

  • Contrastive-based methods excel in time series anomaly detection but rely on computationally expensive attention mechanisms.
  • Existing approaches face challenges in efficiency and framework generality due to complex attention modules.

Purpose of the Study:

  • To develop an attention-free, efficient, and effective solution for contrastive-based time series anomaly detection.
  • To demonstrate that contrastive discrepancy learning can be achieved without attention mechanisms.

Main Methods:

  • Proposed AnomalyTCN, a lightweight, pure convolutional neural network architecture.
  • Designed an asymmetric dual-branch convolution structure and an asymmetric supervision training strategy.
  • Implemented an attention-free contrastive learning framework for anomaly detection.
Keywords:
Contrastive-based time series anomaly detectionDeep learningEfficient time series anomaly detectionPure convolution structureTime series anomaly detection

Related Experiment Videos

Main Results:

  • AnomalyTCN achieved state-of-the-art performance on various time series anomaly detection tasks.
  • Demonstrated significant efficiency gains, saving 83.6% running time and 20.1% memory usage.
  • Validated the effectiveness of a pure convolutional structure for contrastive representation learning.

Conclusions:

  • AnomalyTCN offers a superior balance of performance and efficiency for time series anomaly detection.
  • The study highlights the potential of integrating contrastive learning with efficient time series backbones beyond attention.
  • Paved the way for more accessible and scalable contrastive anomaly detection solutions.