Simple algorithm for recurrent neural networks that can learn sequence completion

István Szita, A. Lőrincz

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We can memorize long sequences like melodies or poems and it is intriguing to develop efficient connectionist representations for this problem. Recurrent neural networks have been proved to offer a reasonable approach here. We start from a few axiomatic assumptions and provide a simple mathematical framework that encapsulates the problem. A gradient-descent based algorithm is derived in this framework. Demonstrations on a benchmark problem show the applicability of our approach.

Original languageEnglish
Title of host publicationIEEE International Conference on Neural Networks - Conference Proceedings
Pages183-188
Number of pages6
Volume1
Publication statusPublished - 2004
Event2004 IEEE International Joint Conference on Neural Networks - Proceedings - Budapest, Hungary
Duration: Jul 25 2004Jul 29 2004

Other

Other2004 IEEE International Joint Conference on Neural Networks - Proceedings
CountryHungary
CityBudapest
Period7/25/047/29/04

Fingerprint

Recurrent neural networks
Demonstrations

ASJC Scopus subject areas

  • Software

Cite this

Szita, I., & Lőrincz, A. (2004). Simple algorithm for recurrent neural networks that can learn sequence completion. In IEEE International Conference on Neural Networks - Conference Proceedings (Vol. 1, pp. 183-188)

Simple algorithm for recurrent neural networks that can learn sequence completion. / Szita, István; Lőrincz, A.

IEEE International Conference on Neural Networks - Conference Proceedings. Vol. 1 2004. p. 183-188.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Szita, I & Lőrincz, A 2004, Simple algorithm for recurrent neural networks that can learn sequence completion. in IEEE International Conference on Neural Networks - Conference Proceedings. vol. 1, pp. 183-188, 2004 IEEE International Joint Conference on Neural Networks - Proceedings, Budapest, Hungary, 7/25/04.
Szita I, Lőrincz A. Simple algorithm for recurrent neural networks that can learn sequence completion. In IEEE International Conference on Neural Networks - Conference Proceedings. Vol. 1. 2004. p. 183-188
Szita, István ; Lőrincz, A. / Simple algorithm for recurrent neural networks that can learn sequence completion. IEEE International Conference on Neural Networks - Conference Proceedings. Vol. 1 2004. pp. 183-188
@inproceedings{11259637cfcb4000bfcbfe3eb0edc479,
title = "Simple algorithm for recurrent neural networks that can learn sequence completion",
abstract = "We can memorize long sequences like melodies or poems and it is intriguing to develop efficient connectionist representations for this problem. Recurrent neural networks have been proved to offer a reasonable approach here. We start from a few axiomatic assumptions and provide a simple mathematical framework that encapsulates the problem. A gradient-descent based algorithm is derived in this framework. Demonstrations on a benchmark problem show the applicability of our approach.",
author = "Istv{\'a}n Szita and A. Lőrincz",
year = "2004",
language = "English",
volume = "1",
pages = "183--188",
booktitle = "IEEE International Conference on Neural Networks - Conference Proceedings",

}

TY - GEN

T1 - Simple algorithm for recurrent neural networks that can learn sequence completion

AU - Szita, István

AU - Lőrincz, A.

PY - 2004

Y1 - 2004

N2 - We can memorize long sequences like melodies or poems and it is intriguing to develop efficient connectionist representations for this problem. Recurrent neural networks have been proved to offer a reasonable approach here. We start from a few axiomatic assumptions and provide a simple mathematical framework that encapsulates the problem. A gradient-descent based algorithm is derived in this framework. Demonstrations on a benchmark problem show the applicability of our approach.

AB - We can memorize long sequences like melodies or poems and it is intriguing to develop efficient connectionist representations for this problem. Recurrent neural networks have been proved to offer a reasonable approach here. We start from a few axiomatic assumptions and provide a simple mathematical framework that encapsulates the problem. A gradient-descent based algorithm is derived in this framework. Demonstrations on a benchmark problem show the applicability of our approach.

UR - http://www.scopus.com/inward/record.url?scp=10944220795&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=10944220795&partnerID=8YFLogxK

M3 - Conference contribution

VL - 1

SP - 183

EP - 188

BT - IEEE International Conference on Neural Networks - Conference Proceedings

ER -