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

Convergence in discrete-time neural networks with specific performance.

T Chu1

  • 1Center for Systems and Control, Department of Mechanics and Engineering Science, Peking University, Beijing 100871, People's Republic of China. tgchu@mech.pku.edu.cn

Physical Review. E, Statistical, Nonlinear, and Soft Matter Physics
|June 21, 2001
PubMed
Summary

This study presents a novel method for analyzing discrete-time neural network convergence. We establish conditions for componentwise exponential stability, enabling the design of networks with predictable performance and decay rates.

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

[Construction of competitive endogenous RNA regulatory network and clinical verification for repeated implantation failure].

Zhonghua yi xue za zhi·2024
Same author

Bioenergetic stress potentiates antimicrobial resistance and persistence.

bioRxiv : the preprint server for biology·2024
Same author

[The Development of SpCas9 Variants with High Specificity and Efficiency Based on the HH Theory].

Molekuliarnaia biologiia·2024
Same author

The Relationship between Psychological Distress and Physical Frailty in Japanese Community-Dwelling Older Adults: A Cross-Sectional Study.

The Journal of frailty & aging·2023
Same author

[The influence of various myelosuppression degrees during neoadjuvant chemotherapy on the curative effect and prognosis of triple-negative breast cancer].

Zhonghua yi xue za zhi·2022
Same author

Equations of motion for weakly compressible point vortices.

Philosophical transactions. Series A, Mathematical, physical, and engineering sciences·2022

Area of Science:

  • Computational Neuroscience
  • Dynamical Systems Theory
  • Machine Learning Theory

Background:

  • Discrete-time neural networks are crucial for complex computations.
  • Analyzing their convergence and stability is essential for reliable performance.
  • Existing methods often struggle with non-symmetric connectivity patterns.

Purpose of the Study:

  • To develop a framework for analyzing convergence in discrete-time neural networks.
  • To establish necessary and sufficient conditions for componentwise absolute (exponential) stability.
  • To enable the design of convergent networks with prescribed performance metrics like decay rate and trajectory bounds.

Main Methods:

  • Decomposition of competitive-cooperative connectivity into cooperative dynamical systems.

Related Experiment Videos

  • Analysis based on order-preserving properties of cooperative systems.
  • Explicit division of connection weights into inhibitory and excitatory components.

    • Main Results:

    • Presentation of simple, necessary, and sufficient conditions for stability and positive invariance.
    • Demonstration of how to design convergent networks with specific performance.
    • Relating competitive-cooperative networks to cooperative dynamical systems without assuming matrix symmetry.

    Conclusions:

    • The proposed decomposition facilitates the analysis of complex neural network dynamics.
    • The derived conditions allow for predictable network performance and stability.
    • This approach bridges formal neural network models with neurophysiological concepts through weight decomposition.