Presentation 1998/3/13
A Study of Back Propagation Learning with a Time-dependent Periodic Activation Function
Yoshitaka TOMONAGA, Masahiro NAKAGAWA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this report, we shall investigate the learning ability of a time-dependent periodic activation function applied to a back propagation learning model. The time-dependent periodic activation function under consideration varies from the sine function to the rectangular one with increasing time. In practice we construct a 3-layered neural network model to learn the XOR and 7-point problems. From the simulation result, the range of learning coefficient with 100% success rate in the present model is found to be wider than sinusoidal function, and the mean of learning epochs is less than the latter model. It is found that the dependence on learning coefficient is to be reduced in proposed model.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Time-dependent Periodic Activation Function / Chaos / Back Propagation
Paper #
Date of Issue

Conference Information
Committee NLP
Conference Date 1998/3/13(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Nonlinear Problems (NLP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Study of Back Propagation Learning with a Time-dependent Periodic Activation Function
Sub Title (in English)
Keyword(1) Time-dependent Periodic Activation Function
Keyword(2) Chaos
Keyword(3) Back Propagation
1st Author's Name Yoshitaka TOMONAGA
1st Author's Affiliation Department of Electrical Engineering, Faculty of Engineering, Nagaoka University of Technology()
2nd Author's Name Masahiro NAKAGAWA
2nd Author's Affiliation Department of Electrical Engineering, Faculty of Engineering, Nagaoka University of Technology
Date 1998/3/13
Paper #
Volume (vol) vol.97
Number (no) 592
Page pp.pp.-
#Pages 8
Date of Issue