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

Sequential monte carlo methods To train neural network models

de Freitas JF1, M Niranjan M, Gee

  • 1Cambridge University Engineering Department, Cambridge CB2 1PZ, U.K.

Neural Computation
|April 19, 2000
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

The Correlation Functions of a Suspension of Large Particles in Amorphous Polybutadiene: A Molecular Dynamics Simulation Study.

Journal of colloid and interface science·2000
Same author

Ring perception: proof of a formula calculating the number of the smallest rings in connected graphs

Journal of chemical information and computer sciences·2000
Same author

Hierarchical bayesian models for regularisation in sequential learning

Neural computation·2000
Same author

Cerebral blood-flow changes evoked by two levels of painful heat stimulation: a positron emission tomography study in humans.

European journal of pain (London, England)·2000
Same author

Water content, raffinose, and dehydrins in the induction of desiccation tolerance in immature wheat embryos

Plant physiology·1999
Same author

Recovery of the flight system following ablation of the tegulae in immature adult locusts

The Journal of experimental biology·1996
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

We introduce a new hybrid algorithm (HySIR) for training neural networks, improving computational speed and accuracy over traditional methods. This global optimization strategy is ideal for complex signal processing and financial market analysis.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computational Finance

Background:

  • Sequential Monte Carlo (SMC) methods are computationally intensive for training neural networks.
  • Existing algorithms often struggle with nonlinear and non-Gaussian signal processing tasks.

Purpose of the Study:

  • To develop a novel, efficient algorithm for training neural networks.
  • To enhance the performance of sequential Monte Carlo techniques in complex environments.
  • To apply the new algorithm to financial market problems, specifically option pricing.

Main Methods:

  • Proposed a hybrid gradient descent sampling importance resampling (HySIR) algorithm.
  • Utilized a sequential framework for learning probability distributions of network weights and outputs.

Related Experiment Videos

  • Compared HySIR against conventional SMC techniques and the extended Kalman filter.

    • Main Results:

    • HySIR demonstrated significant improvements in computational time and accuracy compared to traditional SMC methods.
    • The algorithm effectively handles on-line, nonlinear, and non-Gaussian signal processing.
    • HySIR outperformed the extended Kalman filter in tested problems, including option pricing.

    Conclusions:

    • The HySIR algorithm offers a superior approach to neural network training within a sequential Monte Carlo framework.
    • This method provides a robust global optimization strategy for learning complex probability distributions.
    • HySIR shows promise for advanced applications in signal processing and financial modeling, such as estimating option price densities.