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: Jan 15, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

1.0K

Theoretical Convergence Analysis and Initialization Comparisons of Deep Soft-Thresholding Networks.

Chunyan Xiong, Mengxue Zhang, Tong Wei

    IEEE Transactions on Neural Networks and Learning Systems
    |October 6, 2025
    PubMed
    Summary

    Related Concept Videos

    Difference from Background: Limit of Detection01:05

    Difference from Background: Limit of Detection

    8.0K
    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...
    8.0K

    You might also read

    Related Articles

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

    Sort by
    Same author

    Multi-omics deep learning improves FDG PET-CT-based long-term prognostication of breast cancer.

    NPJ precision oncology·2026
    Same author

    Effects of prone positioning on oxygenation and survival in obese patients with severe acute respiratory distress syndrome: A retrospective cohort study.

    Medicine·2026
    Same author

    Tripodal bisphenolate-bisphosphine ligands with hybrid hard/soft donors: titanium complexes and dinitrogen activation.

    Dalton transactions (Cambridge, England : 2003)·2025
    Same author

    Performance of Solid Oxide Fuel Cells Based on Liquid Hydrocarbon Fuel Reforming Gas: Effect of Cell Structure and Gas Composition.

    ACS applied materials & interfaces·2025
    Same author

    Boost Electrocatalytic Activity of La<sub>0.6</sub>Sr<sub>0.4</sub>Co<sub>0.2</sub>Fe<sub>0.8</sub>O<sub>3-δ</sub> Air Electrode Prepared by High-Temperature Shock for Solid Oxide Electrochemical Cells.

    ACS applied materials & interfaces·2024
    Same author

    In Situ Electrochemical Recovery: Sediment Transformation under Chromium Poisoning in Reversible Solid Oxide Cells with La<sub>0.6</sub>Sr<sub>0.4</sub>Co<sub>0.2</sub>Fe<sub>0.8</sub>O<sub>3-σ</sub>-Based Oxygen Electrodes.

    ACS applied materials & interfaces·2024

    Initializing deep neural networks with soft-thresholding (ST) using identity matrix weights and zero biases accelerates convergence. This method overcomes training challenges like gradient explosion in deep soft-thresholding fully connected networks (ST-FCNs).

    Area of Science:

    • Deep Learning
    • Neural Network Architectures
    • Computational Neuroscience

    Background:

    • Soft-thresholding (ST) is integral to deep neural networks, particularly deep soft-thresholding fully connected networks (ST-FCNs).
    • Training deep ST-FCNs often faces convergence issues, including time-consuming training and gradient explosion, due to an incomplete understanding of their convergence behavior.

    Purpose of the Study:

    • To establish a theoretical link between the convergence of deep ST-FCNs and their weight and bias parameters.
    • To propose an effective initialization strategy for enhancing ST-FCN convergence.

    Main Methods:

    • Theoretical analysis to determine the convergence conditions for deep ST-FCNs as the number of layers approaches infinity.
    • Empirical evaluation comparing identity matrix initialization against standard methods (Gaussian, He, LeCun, Xavier, Uniform) on synthetic and real-world datasets.

    Related Experiment Videos

    Last Updated: Jan 15, 2026

    Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
    03:31

    Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

    Published on: December 15, 2023

    1.0K
  • Validation across various network depths, architectures (DenseNet, ResNet, VGG), and challenging benchmarks.

    • Main Results:

    • Theoretical analysis indicates deep ST-FCNs converge when weights approach an identity matrix and biases approach zero.
    • Identity matrix initialization for weights and zero for biases demonstrated significantly faster and more stable convergence compared to other methods.
    • Findings were robust across diverse datasets (MNIST, CIFAR-10/100, STL-10, Tiny ImageNet) and network complexities.

    Conclusions:

    • Initializing deep ST-FCNs with identity matrix weights and zero biases provides a theoretically grounded and empirically validated method for rapid and stable convergence.
    • This research offers a foundational understanding of ST neural network convergence and extends to recurrent neural networks (RNNs).