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

Universal approximation of multiple nonlinear operators by neural networks.

Andrew D Back1, Tianping Chen

  • 1Windale Technologies, Brisbane, QLD 4075, Australia. back@windale.com

Neural Computation
|November 16, 2002
PubMed
Summary

Neural networks with fixed weights can model multiple nonlinear dynamical systems. This study proposes a theory explaining the mechanism behind this observed phenomenon in artificial intelligence research.

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

Exploration of the Pathogenic Mechanism of the Factor XIII A Subunit in a Patient With Congenital Factor XIII Deficiency.

Molecular genetics & genomic medicine·2026
Same author

Correction: EZH2 suppresses the chemosensitivity of diffuse large B cell lymphoma via regulating TP53INP1.

Clinical and experimental medicine·2026
Same author

EZH2 suppresses the chemosensitivity of diffuse large B cell lymphoma via regulating TP53INP1.

Clinical and experimental medicine·2025
Same author

Development of an HPLC method for the determination of amines in a leukemia mouse model.

Analytical methods : advancing methods and applications·2025
Same author

Successful treatment of iodixanol-induced type II Kounis syndrome with an intra-aortic balloon pump: a case report and literature review.

Frontiers in cardiovascular medicine·2025
Same author

Modulation of copper homeostasis and cuproptosis by PDHA1 in acute myeloid leukemia.

Discover oncology·2025

Area of Science:

  • Artificial Intelligence
  • Computational Neuroscience
  • Dynamical Systems Theory

Background:

  • Recent research highlights the capacity of certain neural network architectures with static, unchanging weights to simulate diverse nonlinear dynamical systems.
  • Despite observational evidence from computational simulations, the underlying theoretical framework explaining this multi-system modeling capability remains incompletely understood.

Purpose of the Study:

  • To propose a theoretical explanation for how neural networks with fixed weights can effectively model multiple distinct nonlinear dynamical systems.
  • To elucidate the potential mechanisms enabling this versatile computational property.

Main Methods:

  • Theoretical analysis of neural network dynamics.
  • Mathematical modeling of information processing in fixed-weight networks.

Related Experiment Videos

Main Results:

  • A theoretical framework is presented that elucidates a potential mechanism for multiple nonlinear dynamical system modeling by fixed-weight neural networks.
  • The proposed theory offers insights into the computational principles underlying this observed phenomenon.

Conclusions:

  • The study provides a theoretical foundation for understanding the multi-system modeling capabilities of fixed-weight neural networks.
  • This work contributes to the theoretical understanding of neural network function in modeling complex dynamics.