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

Continuous-time symmetric Hopfield nets are computationally universal.

Jirí Síma1, Pekka Orponen

  • 1Institute of Computer Science, Academy of Sciences of the Czech Republic, P.O. Box 5 182 07 Prague 8, Czech Republic. sima@cs.cas.cz

Neural Computation
|March 7, 2003
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

Automated design of scaffold-free DNA wireframe nanostructures.

Nature communications·2025
Same author

On energy complexity of fully-connected layers.

Neural networks : the official journal of the International Neural Network Society·2024
Same author

Energy Complexity of Convolutional Neural Networks.

Neural computation·2024
Same author

DNAforge: a design tool for nucleic acid wireframe nanostructures.

Nucleic acids research·2024
Same author

Algorithmic Design of 3D Wireframe RNA Polyhedra.

ACS nano·2022
Same author

RNA secondary structure prediction with convolutional neural networks.

BMC bioinformatics·2022

Continuous-time symmetric Hopfield networks are proven to be universal and efficient computational devices. These networks can simulate general computations performed by discrete-time recurrent neural networks, demonstrating their power in computation theory.

Area of Science:

  • Computational neuroscience
  • Theory of computation
  • Dynamical systems

Background:

  • Continuous-time symmetric Hopfield networks possess constrained dynamics controlled by Lyapunov functions.
  • Understanding the computational capabilities of these networks is crucial for advancing artificial intelligence and computational theory.

Purpose of the Study:

  • To establish that continuous-time symmetric Hopfield networks are capable of general computation.
  • To demonstrate their universality and efficiency as computational devices.

Main Methods:

  • Simulating convergent synchronous fully parallel computations of discrete-time recurrent neural networks using symmetric continuous-time Hopfield nets.
  • Analyzing the relationship between discrete-time network parameters and continuous-time Hopfield net operation time.

Related Experiment Videos

Main Results:

  • Any convergent synchronous fully parallel computation by a discrete-time recurrent neural network can be simulated by a symmetric continuous-time Hopfield net with 18n + 7 units.
  • The simulation operates in continuous time Theta(t*/epsilon), dependent on the discrete-time network's maximum weight size and convergence time.
  • This implies that polynomially space-bounded Turing machines can be simulated by polynomial-size continuous-time symmetric Hopfield nets.

Conclusions:

  • Continuous-time symmetric Hopfield networks are universal and efficient computational devices.
  • These findings expand the understanding of computation within dynamical systems.
  • The results have implications for the design and capabilities of neural network models in computation theory.