Presentation 1994/5/20
On Patten presentation for learning One-demensional discrete dynamical systems in neural networks
Hiroyuki Nakajima, Takafumi / Ohnishi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) How the restriction on the size of training set and modification of distribution density and time correlation affect thd speed and accuracy of learning discrete dynamical systems in neural networks are studied by numerical experiments on learning Cubic map and Logistic map.These systems can be learned with about 20 training data as accurately as learning with infinite data.Acceleration of learning by trajectory error method is also observed.The distribution density of patterns plays a more important role in learning than time correlation.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) neural network / back propagation / discrete dynamical system / chaos / invariant measure
Paper # NLP94-13
Date of Issue

Conference Information
Committee NLP
Conference Date 1994/5/20(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) On Patten presentation for learning One-demensional discrete dynamical systems in neural networks
Sub Title (in English)
Keyword(1) neural network
Keyword(2) back propagation
Keyword(3) discrete dynamical system
Keyword(4) chaos
Keyword(5) invariant measure
1st Author's Name Hiroyuki Nakajima
1st Author's Affiliation Department of Electrical Engineering II,Faculty of Engineering Kyoto University()
2nd Author's Name Takafumi / Ohnishi
2nd Author's Affiliation Department of Electrical Engineering II,Faculty of Engineering Kyoto University
Date 1994/5/20
Paper # NLP94-13
Volume (vol) vol.94
Number (no) 45
Page pp.pp.-
#Pages 8
Date of Issue