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

Density estimation and random variate generation using multilayer networks.

M Magdon-Ismail1, A Atiya

  • 1Dept. of Comput. Sci., Rensselaer Polytech. Inst., Troy, NY.

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

Prevalence and characteristics of cigarillo smoking in Canada: results from the Canadian Tobacco Use Monitoring Survey.

Public health·2018
Same author

The equivalent martingale measure: an introduction to pricing using expectations.

IEEE transactions on neural networks·2008
Same author

An algorithmic approach to adaptive state filtering using recurrent neural networks.

IEEE transactions on neural networks·2008
Same author

The early restart algorithm.

Neural computation·2000
Same author

No free lunch for noise prediction.

Neural computation·2000
Same author

No free lunch for early stopping.

Neural computation·1999
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 paper introduces novel neural network methods for density estimation and random variate generation, achieving near-optimal convergence rates for improved data analysis and simulation.

Area of Science:

  • Computational Statistics
  • Machine Learning
  • Numerical Analysis

Background:

  • Density estimation and random variate generation are fundamental in statistical modeling and simulation.
  • Existing methods often have limitations regarding the types of distributions they can handle or their convergence rates.
  • Multilayer neural networks offer a flexible framework for complex function approximation.

Purpose of the Study:

  • To develop new, efficient methods for density estimation using neural networks.
  • To introduce novel techniques for random variate generation applicable to a wide range of densities.
  • To provide theoretical convergence guarantees for the proposed methods.

Main Methods:

  • Developed two novel neural network-based methods for density estimation: a stochastic and a deterministic approach, by approximating the distribution function.

Related Experiment Videos

  • Introduced two new random variate generation methods: one leveraging an inverse relationship with density estimation (stochastic and deterministic variants), and another using a control formulation with a 'controller network'.
  • Utilized multilayer neural networks for implementing both density estimation and random variate generation frameworks.

    • Main Results:

    • Proved theoretical convergence results, demonstrating L(infinity) convergence rates of O((log log N/N)((1-epsilon)/2)) for both density estimation and random variate generation.
    • Showcased that these convergence rates are near-optimal for sufficiently smooth target densities, outperforming traditional kernel density estimators in certain aspects.
    • Numerical simulations confirmed the practical performance and effectiveness of the proposed neural network-based methods.

    Conclusions:

    • The proposed neural network framework provides a versatile and powerful approach to density estimation and random variate generation.
    • The methods overcome limitations of existing techniques and offer strong theoretical convergence guarantees.
    • The study offers a valuable reference for practitioners through an extended introduction and bibliography.