SIMD型並列計算機を用いたカオス誤差逆伝播学習モデル

概要

論文の詳細を見る
In this study, an error black propagation scheme with periodic chaos neuron model is proposed using the recently developed neuro-computer, MY-NEUPOWER. This learning model is properly characterized in terms of periodic chaos neurons to involve chaotic dynamics as well as external control of the periodicity. MY-NEUPOWER is found to increase the learning speed of a neuron model. In practice, a 3 -layered neural network model is applied to the n-bit parity check problem and 7-point problem. From the present simulation results, the learning speed with MY-NEUPOWER is found to be faster than a serial computer at relatively large bit parity problems.
長岡技術科学大学の論文
2000-03-29