Opposite transfer functions and backpropagation through time

Mario Ventresca; Hamid R. Tizhoosh

doi:10.1109/FOCI.2007.371529

Opposite transfer functions and backpropagation through time

Mario Ventresca, Hamid R. Tizhoosh

Artificial Intelligence and Informatics

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Backpropagation through time is a very popular discrete-time recurrent neural network training algorithm. However, the computational time associated with the learning process to achieve high accuracy is high. While many approaches have been proposed that alter the learning algorithm, this paper presents a computationally inexpensive method based on the concept of opposite transfer functions to improve learning in the backpropagation through time algorithm. Specifically, we will show an improvement in the accuracy, stability as well as an acceleration in learning time. We will utilize three common benchmarks to provide experimental evidence of the improvements.

Original language	English (US)
Title of host publication	Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007
Pages	570-577
Number of pages	8
DOIs	https://doi.org/10.1109/FOCI.2007.371529
State	Published - 2007
Event	2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007 - Honolulu, HI, United States Duration: Apr 1 2007 → Apr 5 2007

Publication series

Name	Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007

Conference

Conference	2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007
Country/Territory	United States
City	Honolulu, HI
Period	4/1/07 → 4/5/07

Keywords

Backpropagation through time
Opposite transfer functions
Opposition-based learning

ASJC Scopus subject areas

Artificial Intelligence
Software
General Mathematics

Access to Document

10.1109/FOCI.2007.371529

Cite this

Ventresca, M., & Tizhoosh, H. R. (2007). Opposite transfer functions and backpropagation through time. In Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007 (pp. 570-577). Article 4233963 (Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007). https://doi.org/10.1109/FOCI.2007.371529

Opposite transfer functions and backpropagation through time. / Ventresca, Mario; Tizhoosh, Hamid R.
Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007. 2007. p. 570-577 4233963 (Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Ventresca, M & Tizhoosh, HR 2007, Opposite transfer functions and backpropagation through time. in Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007., 4233963, Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007, pp. 570-577, 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007, Honolulu, HI, United States, 4/1/07. https://doi.org/10.1109/FOCI.2007.371529

@inproceedings{6a03977173fa4cf58e8753bfef21042a,

title = "Opposite transfer functions and backpropagation through time",

abstract = "Backpropagation through time is a very popular discrete-time recurrent neural network training algorithm. However, the computational time associated with the learning process to achieve high accuracy is high. While many approaches have been proposed that alter the learning algorithm, this paper presents a computationally inexpensive method based on the concept of opposite transfer functions to improve learning in the backpropagation through time algorithm. Specifically, we will show an improvement in the accuracy, stability as well as an acceleration in learning time. We will utilize three common benchmarks to provide experimental evidence of the improvements.",

keywords = "Backpropagation through time, Opposite transfer functions, Opposition-based learning",

author = "Mario Ventresca and Tizhoosh, {Hamid R.}",

year = "2007",

doi = "10.1109/FOCI.2007.371529",

language = "English (US)",

isbn = "1424407036",

series = "Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007",

pages = "570--577",

booktitle = "Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007",

note = "2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007 ; Conference date: 01-04-2007 Through 05-04-2007",

}

TY - GEN

T1 - Opposite transfer functions and backpropagation through time

AU - Ventresca, Mario

AU - Tizhoosh, Hamid R.

PY - 2007

Y1 - 2007

N2 - Backpropagation through time is a very popular discrete-time recurrent neural network training algorithm. However, the computational time associated with the learning process to achieve high accuracy is high. While many approaches have been proposed that alter the learning algorithm, this paper presents a computationally inexpensive method based on the concept of opposite transfer functions to improve learning in the backpropagation through time algorithm. Specifically, we will show an improvement in the accuracy, stability as well as an acceleration in learning time. We will utilize three common benchmarks to provide experimental evidence of the improvements.

AB - Backpropagation through time is a very popular discrete-time recurrent neural network training algorithm. However, the computational time associated with the learning process to achieve high accuracy is high. While many approaches have been proposed that alter the learning algorithm, this paper presents a computationally inexpensive method based on the concept of opposite transfer functions to improve learning in the backpropagation through time algorithm. Specifically, we will show an improvement in the accuracy, stability as well as an acceleration in learning time. We will utilize three common benchmarks to provide experimental evidence of the improvements.

KW - Backpropagation through time

KW - Opposite transfer functions

KW - Opposition-based learning

UR - http://www.scopus.com/inward/record.url?scp=34548819318&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34548819318&partnerID=8YFLogxK

U2 - 10.1109/FOCI.2007.371529

DO - 10.1109/FOCI.2007.371529

M3 - Conference contribution

AN - SCOPUS:34548819318

SN - 1424407036

SN - 9781424407033

T3 - Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007

SP - 570

EP - 577

BT - Proceedings of the 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007

T2 - 2007 IEEE Symposium on Foundations of Computational Intelligence, FOCI 2007

Y2 - 1 April 2007 through 5 April 2007

ER -

Opposite transfer functions and backpropagation through time

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this