講演抄録/キーワード |
講演名 |
2009-11-13 10:15
A Novel Chaos Associative Memory with Tchebycheff Activation Function ○Masahiro Nakagawa(Nagaoka Univ. of Tech.) NLP2009-101 |
抄録 |
(和) |
In this presentation we shall put forward a novel chaos neuron model and investigate the dynamic properties. The present artificial
neuron model is defined by such periodic input-output mapping as the Tchebycheff function of the second kind. It is apparently shown that the present
neural network with a periodic activation function has an ability of the retrievals of the embedded patterns superior than the conventional neural network
with such a monotonous mapping. This advantage is considered to be as a result of the nonmonotonous property of the periodic mapping which
involves a chaotic behaviour of the neurons. It is also found that the present chaos neuron model has a remarkably larger memory capacity than the
conventional association system with the monotonous dynamics. These findings are considered to result from the chaotic dynamics to avoid at an
unfavourable spurious states. |
(英) |
In this presentation we shall put forward a novel chaos neuron model and investigate the dynamic properties. The present artificial
neuron model is defined by such periodic input-output mapping as the Tchebycheff function of the second kind. It is apparently shown that the present
neural network with a periodic activation function has an ability of the retrievals of the embedded patterns superior than the conventional neural network
with such a monotonous mapping. This advantage is considered to be as a result of the nonmonotonous property of the periodic mapping which
involves a chaotic behaviour of the neurons. It is also found that the present chaos neuron model has a remarkably larger memory capacity than the
conventional association system with the monotonous dynamics. These findings are considered to result from the chaotic dynamics to avoid at an
unfavourable spurious states. |
キーワード |
(和) |
chaos neuron / periodic mapping, / ,associative memory / / / / / |
(英) |
/ / / / / / / |
文献情報 |
信学技報, vol. 109, no. 269, NLP2009-101, pp. 109-114, 2009年11月. |
資料番号 |
NLP2009-101 |
発行日 |
2009-11-04 (NLP) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
NLP2009-101 |