Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

Optimization-based learning with bounded error for feedforward neural networks.

A Alessandri¹, M Sanguineti, M Maggiore

¹Naval Autom. Inst., Nat. Res. Council of Italy, Genoa.

IEEE Transactions on Neural Networks

|February 5, 2008

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

Synergies of THESEUS with the large facilities of the 2030s and guest observer opportunities.

Experimental astronomy·2022

Same author

Development and installation of a radio frequency quadrupole cooler test.

The Review of scientific instruments·2019

Same author

The cosmic ray shadow of the Moon observed with the ANTARES neutrino telescope.

The European physical journal. C, Particles and fields·2019

Same author

An algorithm for the reconstruction of high-energy neutrino-induced particle showers and its application to the ANTARES neutrino telescope.

The European physical journal. C, Particles and fields·2017

Same author

Sperm whale long-range echolocation sounds revealed by ANTARES, a deep-sea neutrino telescope.

Scientific reports·2017

Same author

Integration of RFQ beam coolers and solenoidal magnetic fields.

The Review of scientific instruments·2016

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

A novel optimization algorithm efficiently trains feedforward neural networks using a sliding-window cost. This method proves effective for large datasets and outperforms traditional backpropagation and extended Kalman filter learning approaches.

Area of Science:

Artificial Intelligence
Machine Learning
Computational Neuroscience

Background:

Feedforward neural networks (FNNs) are fundamental in machine learning.
Efficient training algorithms are crucial for handling large datasets.
Existing methods like backpropagation and extended Kalman filter have limitations.

Purpose of the Study:

To introduce a new optimization-based learning algorithm for FNNs.
To address computational efficiency for large-scale data.
To analyze the algorithm's convergence and robustness.

Main Methods:

The proposed algorithm determines network weights by minimizing a sliding-window cost function.
It is designed for batch learning scenarios.
Convergence and robustness properties are mathematically analyzed.

Related Experiment Videos

Main Results:

Simulation results demonstrate the algorithm's effectiveness.
The algorithm shows advantages over backpropagation and extended Kalman filter.
Efficient handling of large datasets is confirmed.

Conclusions:

The presented optimization algorithm offers an effective and efficient approach for training FNNs.
It provides a viable alternative to existing learning methods, especially for large datasets.
The algorithm's robustness and convergence properties are well-established.