On-line learning in multilayer neural networks

David Saad, Sara A. Solla

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

We present an analytic solution to the problem of on-line gradient-descent learning for two-layer neural networks with an arbitrary number of hidden units in both teacher and student networks. The technique, demonstrated here for the case of adaptive input-to-hidden weights, becomes exact as the dimensionality of the input space increases.
Original languageEnglish
Title of host publicationProceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications
Place of PublicationOxford
PublisherKluwer
ISBN (Print)0-7923-99331
DOIs
Publication statusPublished - 1997

Fingerprint

Multilayer neural networks
Students
Neural networks

Bibliographical note

© Springer Science+Business Media New York 1997

Keywords

  • algorithm
  • design
  • measurement
  • performance
  • theory
  • verification

Cite this

Saad, D., & Solla, S. A. (1997). On-line learning in multilayer neural networks. In Proceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications Oxford: Kluwer. https://doi.org/10.1007/978-1-4615-6099-9_53
Saad, David ; Solla, Sara A. / On-line learning in multilayer neural networks. Proceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications. Oxford : Kluwer, 1997.
@inbook{f063d37c0c7b413ab570525d37079350,
title = "On-line learning in multilayer neural networks",
abstract = "We present an analytic solution to the problem of on-line gradient-descent learning for two-layer neural networks with an arbitrary number of hidden units in both teacher and student networks. The technique, demonstrated here for the case of adaptive input-to-hidden weights, becomes exact as the dimensionality of the input space increases.",
keywords = "algorithm, design, measurement, performance, theory, verification",
author = "David Saad and Solla, {Sara A.}",
note = "{\circledC} Springer Science+Business Media New York 1997",
year = "1997",
doi = "10.1007/978-1-4615-6099-9_53",
language = "English",
isbn = "0-7923-99331",
booktitle = "Proceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications",
publisher = "Kluwer",
address = "Netherlands",

}

Saad, D & Solla, SA 1997, On-line learning in multilayer neural networks. in Proceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications. Kluwer, Oxford. https://doi.org/10.1007/978-1-4615-6099-9_53

On-line learning in multilayer neural networks. / Saad, David; Solla, Sara A.

Proceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications. Oxford : Kluwer, 1997.

Research output: Chapter in Book/Report/Conference proceedingChapter

TY - CHAP

T1 - On-line learning in multilayer neural networks

AU - Saad, David

AU - Solla, Sara A.

N1 - © Springer Science+Business Media New York 1997

PY - 1997

Y1 - 1997

N2 - We present an analytic solution to the problem of on-line gradient-descent learning for two-layer neural networks with an arbitrary number of hidden units in both teacher and student networks. The technique, demonstrated here for the case of adaptive input-to-hidden weights, becomes exact as the dimensionality of the input space increases.

AB - We present an analytic solution to the problem of on-line gradient-descent learning for two-layer neural networks with an arbitrary number of hidden units in both teacher and student networks. The technique, demonstrated here for the case of adaptive input-to-hidden weights, becomes exact as the dimensionality of the input space increases.

KW - algorithm

KW - design

KW - measurement

KW - performance

KW - theory

KW - verification

UR - https://link.springer.com/chapter/10.1007%2F978-1-4615-6099-9_53

U2 - 10.1007/978-1-4615-6099-9_53

DO - 10.1007/978-1-4615-6099-9_53

M3 - Chapter

SN - 0-7923-99331

BT - Proceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications

PB - Kluwer

CY - Oxford

ER -

Saad D, Solla SA. On-line learning in multilayer neural networks. In Proceedings of the first international conference on mathematics of neural networks : models, algorithms and applications: models, algorithms and applications. Oxford: Kluwer. 1997 https://doi.org/10.1007/978-1-4615-6099-9_53