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

Performance analysis of a pipelined backpropagation parallel algorithm.

A Petrowski¹, G Dreyfus, C Girault

¹Dept. of Inf., Inst. Nat. des Telecommun., Evry.

IEEE Transactions on Neural Networks

|January 1, 1993

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

[Thoracic electrical impedance tomography: Technical aspects and clinical applications].

Revue des maladies respiratoires·2025

Same author

Promoting large multinational academic clinical trials for gastrointestinal cancers in Europe: the ENGIC (European Network for GI Cancers) initiative.

ESMO open·2025

Same author

[High-flow nasal oxygen therapy and hypercapnic acute respiratory failure].

Revue des maladies respiratoires·2024

Same author

[Ventilatory management of SARS-CoV-2 acute respiratory failure].

Revue des maladies respiratoires·2023

Same author

Revue des maladies respiratoires actualites·2022

Same author

[High-flow nasal oxygen therapy in acute respiratory failure: From physiology to evidence-based medicine!]

Revue des maladies respiratoires·2022

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 an efficient parallel algorithm for training feedforward neural networks. By treating network layers as pipeline stages, the method achieves high performance across various network architectures.

Area of Science:

Artificial Intelligence
Machine Learning
Computer Science

Background:

Supervised training of feedforward neural networks commonly utilizes the error backpropagation algorithm.
Existing parallel training methods may face challenges with efficiency and architectural independence.

Purpose of the Study:

To develop an efficient parallel algorithm for training feedforward neural networks.
To demonstrate that the parallelization's performance is largely independent of network architecture.

Main Methods:

The successive layers of a feedforward neural network are modeled as stages in a pipeline.
A simple placement rule is employed to enable simultaneous computations across network layers.

Main Results:

Related Experiment Videos

Analytic expressions confirm the efficiency of the proposed parallelization strategy.
Implementation performance shows minimal dependence on the specific neural network architecture.
Experimental results align with analytical predictions.

Conclusions:

The described pipeline-based parallel algorithm offers an efficient approach to training feedforward neural networks.
The method's architectural independence simplifies performance prediction on parallel computing systems.
This approach facilitates easier estimation of learning performance for diverse neural network designs on parallel hardware.