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 Video

Updated: Oct 8, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

720

Memory-Augmented Generative Adversarial Networks for Anomaly Detection.

Ziyi Yang, Teng Zhang, Iman Soltani Bozchalooi

    IEEE Transactions on Neural Networks and Learning Systems
    |December 28, 2021
    PubMed
    Summary
    This summary is machine-generated.

    Related Concept Videos

    You might also read

    Related Articles

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

    Sort by
    Same author

    Learning reduced-order models for cardiovascular simulations with graph neural networks.

    Computers in biology and medicine·2023
    Same author

    Nomograms for Predicting the Risk of SIRS and Urosepsis after Uroscopic Minimally Invasive Lithotripsy.

    BioMed research international·2022
    Same author

    The α7 Nicotinic Acetylcholine Receptor Agonist GTS-21 Improves Bacterial Clearance <i>via</i> Regulation of Monocyte Recruitment and Activity in Polymicrobial Septic Peritonitis.

    Frontiers in immunology·2022
    Same author

    Ionically Active MXene Nanopore Actuators.

    Small (Weinheim an der Bergstrasse, Germany)·2022
    Same author

    Global profiling reveals common and distinct N6-methyladenosine (m6A) regulation of innate immune responses during bacterial and viral infections.

    Cell death & disease·2022
    Same author

    Phyllotaxis without symmetry: what can we learn from flower heads?

    Journal of experimental botany·2022

    We introduce MEMGAN, a novel memory-augmented deep learning model for robust semisupervised anomaly detection (AD). MEMGAN enhances traditional methods by using external memory units, improving the reliability of identifying abnormal data.

    Area of Science:

    • Artificial Intelligence
    • Machine Learning
    • Computer Vision

    Background:

    • Traditional anomaly detection (AD) methods often struggle with distinguishing normal from abnormal data due to limitations in modeling data distributions.
    • Existing deep learning approaches require additional constraints for effective anomaly identification.

    Purpose of the Study:

    • To propose a novel memory-augmented deep learning model for semisupervised anomaly detection.
    • To enhance the robustness and reliability of anomaly detection by leveraging external memory units.

    Main Methods:

    • Development of memory augmented generative adversarial networks (MEMGAN), integrating external memory units via attentional operations.
    • Utilizing a property where encoded normal data reside within the convex hull of memory units, while abnormal data fall outside.

    Related Experiment Videos

    Last Updated: Oct 8, 2025

    Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
    03:14

    Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

    Published on: December 6, 2024

    720

    Main Results:

    • MEMGAN demonstrated significant improvements over existing AD models on diverse datasets (MVTec, MNIST, CIFAR10, Arrhythmia).
    • The model exhibits a more robust and reliable anomaly detection process due to its unique latent space properties.
    • Decoded memory units in MEMGAN showed increased diversity and interpretability compared to prior memory-augmented models.

    Conclusions:

    • MEMGAN offers a superior approach to semisupervised anomaly detection by incorporating memory augmentation.
    • The model's design enhances interpretability and diversity in anomaly detection tasks.
    • This work advances the field of anomaly detection with a more effective and reliable deep learning framework.