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

Neural network uncertainty assessment using Bayesian statistics: a remote sensing application.

F Aires1, C Prigent, W B Rossow

  • 1Department of Applied Physics and Applied Mathematics, Columbia University, NASA Goddard Institute for Space Studies, New York, NY 10025, USA. faires@giss.nasa.gov

Neural Computation
|October 13, 2004
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

Clinical Outcomes and Toxicity Profile of Chemoradiotherapy in Older Versus Younger Patients With Anal Cancer: A Retrospective Cohort Analysis.

Clinical oncology (Royal College of Radiologists (Great Britain))·2025
Same author

Results of surgery versus stereotactic body radiotherapy for lung cancer.

Pulmonology·2022
Same author

Low energy Ne ion beam induced-modifications of magnetic properties in MnAs thin films.

Journal of physics. Condensed matter : an Institute of Physics journal·2016
Same author

Efficient broadband 400  nm noncollinear second-harmonic generation of chirped femtosecond laser pulses in BBO and LBO.

Applied optics·2014
Same author

Electronic temperatures, densities, and plasma x-ray emission of a 14.5 GHz electron-cyclotron resonance ion source.

The Review of scientific instruments·2010
Same author

Aerosol retrievals over the ocean by use of channels 1 and 2 AVHRR data: sensitivity analysis and preliminary results.

Applied optics·2008
Same journal

A Model-Free Reinforcement Learning Implementation of Decision Making Under Uncertainty by Sequential Sampling.

Neural computation·2026
Same journal

DROP: Distributional and Regular Optimism and Pessimism for Reinforcement Learning.

Neural computation·2026
Same journal

Hierarchical Active Inference Using Successor Representations.

Neural computation·2026
Same journal

W-Kernel and Its Principal Space for Frequentist Evaluation of Bayesian Estimators.

Neural computation·2026
Same journal

A Hidden Markov Model-Inspired Sequence Classification Method for Hyperdimensional Computing.

Neural computation·2026
Same journal

Sparse Graphical Modeling for Electrophysiological Phase-Based Connectivity Using Circular Statistics.

Neural computation·2026
See all related articles

This study introduces a Bayesian technique for neural network (NN) uncertainty estimation in remote sensing. It quantifies parameter and output uncertainties, enhancing model robustness and physical interpretability for applications like temperature retrieval.

Area of Science:

  • Remote Sensing
  • Machine Learning
  • Geophysics

Background:

  • Neural networks (NNs) are effective for regression in remote sensing but often lack uncertainty estimates.
  • Traditional NN approaches use point estimation, limiting model robustness assessment.
  • Understanding NN internal structure and dependencies is crucial for reliable, physically coherent models.

Purpose of the Study:

  • To present a Bayesian technique for evaluating uncertainties in NN parameters (synaptic weights).
  • To demonstrate the computation of output uncertainties (error bars, error correlations) from weight uncertainties.
  • To analyze uncertainties in NN Jacobians for assessing model internal structure and physical meaning.

Main Methods:

  • Bayesian inference to estimate uncertainties of NN weights.

Related Experiment Videos

  • Analytical computation of output uncertainties based on weight uncertainties.
  • Monte Carlo integration to assess NN Jacobian robustness.
  • Principal component analysis-based regularization for Jacobian stabilization.

    • Main Results:

    • Demonstrated application to retrieve surface skin temperature, microwave emissivities, and water vapor content using satellite data.
    • Weight uncertainty estimates enable monitoring of NN model robustness.
    • Jacobian analysis reveals model dependency structures, improving physical interpretability.

    Conclusions:

    • The proposed Bayesian approach provides crucial uncertainty estimates for NN parameters and outputs in remote sensing.
    • Quantifying Jacobian uncertainties enhances the reliability and physical interpretability of NN regression models.
    • This method moves beyond black-box models towards robust, physically meaningful applications.