Presentation 1995/5/19
Basins of the Optimal States in Learning Dynamical Systems
Hiroyuki Nakajima, Yoshisuke Ueda,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A new phenomenon called a "riddled basin", which is an extremely complicated domain of attraction usually accompanying invariant sets for dynamical systems driven by chaotic signal, e.g. chaotic synchronizing systems, have been drawing attention of nonlinear scientists. In this report, the optimal state of a back-propagation learning of chaotic time series is shown to be a "locally riddled basin" by a numerical experiment. It is also proved that the optimal state of learning the parameter of the tent map by a gradient-descent method can be an attractor with a riddled basin.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) basin / neural network / back-propagation / dynamical system / chaos / gradient-descent method
Paper #
Date of Issue

Conference Information
Committee NLP
Conference Date 1995/5/19(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 ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Basins of the Optimal States in Learning Dynamical Systems
Sub Title (in English)
Keyword(1) basin
Keyword(2) neural network
Keyword(3) back-propagation
Keyword(4) dynamical system
Keyword(5) chaos
Keyword(6) gradient-descent method
1st Author's Name Hiroyuki Nakajima
1st Author's Affiliation Department of Electrical and Electronic Engineering Faculty of Engineering, Kyoto University()
2nd Author's Name Yoshisuke Ueda
2nd Author's Affiliation Department of Electrical and Electronic Engineering Faculty of Engineering, Kyoto University
Date 1995/5/19
Paper #
Volume (vol) vol.95
Number (no) 47
Page pp.pp.-
#Pages 8
Date of Issue