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

Bayesian approach to neural-network modeling with input uncertainty.

W A Wright1

  • 1AIP Department at The Sowerby Research Centre, British Aerospace, Bristol, BS12 7QW, UK.

IEEE Transactions on Neural Networks
|February 7, 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

Tetracycline-microbiologic, pharmacologic, and clinical evaluation.

Antibiotics & chemotherapy (Northfield, Ill.)·2014
Same author

Holographic implementation of a Hopfield model with discrete weightings.

Applied optics·2010
Same author

The general practitioner in his community.

The Journal-lancet·2010
Same author

Development of a measure of patient satisfaction with monitored anesthesia care: the Iowa Satisfaction with Anesthesia Scale.

Anesthesiology·1997
Same author

J-wire versus straight wire for central venous system cannulation via the external jugular vein.

Anesthesia and analgesia·1982
Same author

Vasopressin administration in the first month of life: effects of growth and water metabolism in hypothalamic diabetes insipidus rats.

Pharmacology, biochemistry, and behavior·1977
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 introduces a Bayesian neural network framework to handle noisy input data, a common issue in real-world applications. The method accounts for input noise, improving accuracy and enabling inference on noiseless data.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computational Statistics

Background:

  • Bayesian inference in neural networks typically assumes noise-free input data.
  • Real-world data often suffers from noise or corruption, violating this assumption.
  • Existing methods are inadequate for handling input noise in Bayesian neural networks.

Purpose of the Study:

  • To develop a Bayesian neural network framework capable of handling noisy input data.
  • To provide a robust method for real-world applications with errors-in-variables.
  • To enable accurate inference even when input data is corrupted.

Main Methods:

  • Developed a Bayesian neural network framework incorporating a noise model for input data.
  • Applied Laplace approximation to analyze the impact of small, symmetric noise processes.

Related Experiment Videos

  • Utilized Markov chain Monte Carlo (MCMC) methods to sample network weights and noiseless inputs jointly.

    • Main Results:

    • Introduced an additional term to Bayesian error bars, dependent on input noise variance.
    • Demonstrated the ability to infer regression over the true, noiseless input data.
    • Showcased the framework's effectiveness in handling corrupted input data.

    Conclusions:

    • The proposed Bayesian neural network framework effectively accommodates input noise.
    • This approach enhances the reliability of Bayesian neural networks for real-world problems.
    • The method allows for the recovery of underlying noiseless data patterns.