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

Gaussian activation functions using Markov chains.

H C Card1, D K McNeill

  • 1Dept. of Electr. and Comput. Eng., Manitoba Univ., Winnipeg, Man., Canada.

IEEE Transactions on Neural Networks
|February 5, 2008
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

Near IR interband transitions and optical parameters of metal-germanium contacts.

Applied optics·2010
Same author

Photocounting distributions for exponentially decaying sources.

Optics letters·2009
Same author

Doubly stochastic Poisson processes in artificial neural learning.

IEEE transactions on neural networks·2008
Same author

Vector quantization of images using modified adaptive resonance algorithm for hierarchical clustering.

IEEE transactions on neural networks·2008
Same author

Compound binomial processes in neural integration.

IEEE transactions on neural networks·2008
Same author

Stochastic radial basis functions.

International journal of neural systems·2003
Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013
Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011
Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011
Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011
Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011
Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011
See all related articles

This study enhances stochastic signal processing (SSP) for neural networks by analyzing signal-to-noise limits and generalizing stochastic counters for Gaussian mixture models, quantifying hardware trade-offs.

Area of Science:

  • Computer Science
  • Electrical Engineering
  • Artificial Intelligence

Background:

  • Earlier work implemented sigmoidal neural nonlinearities using stochastic counters.
  • Stochastic signal processing (SSP) offers potential hardware advantages for neural networks.

Purpose of the Study:

  • To extend prior research on stochastic counters for neural network implementations.
  • To define signal-to-noise limitations in stochastic arithmetic and signal processing.
  • To generalize stochastic counters for Gaussian mixture models and analyze hardware trade-offs.

Main Methods:

  • Analysis of unipolar and bipolar stochastic arithmetic and signal processing.
  • Generalization of stochastic counters to include neural transfer functions for Gaussian mixture models.

Related Experiment Videos

  • Quantitative analysis of processing time versus hardware advantages for SSP.
  • Evaluation of Gaussian activation function realization in pulsed digital networks using stochastic signals.

    • Main Results:

    • Defined signal-to-noise limitations for unipolar and bipolar stochastic signal processing.
    • Generalized stochastic counters to accommodate neural transfer functions in Gaussian mixture models.
    • Quantified the trade-off between hardware advantages and potential processing time increases in SSP.
    • Analyzed the quantitative feasibility of accurate Gaussian activation functions using stochastic signals in digital networks.

    Conclusions:

    • Stochastic counters can be generalized for advanced neural network components like Gaussian mixture models.
    • The efficiency of stochastic signal processing in hardware requires careful consideration of processing time.
    • Accurate Gaussian activation functions are achievable in pulsed digital networks using stochastic signal processing.