生成学習アルゴリズムにおける構造化RNNのダイナミックスの学習

概要

論文の詳細を見る
In this report,a. new learning algorithm called constructive learning algorithm and a module structured recurrent neural network are proposed. The network has the ability to memorize and regenerate any wave-forms. In particular, this report shows any periodical waveforms can be approximated by using the minimum number of elementary modules. This network is constructed by adaptive oscillatimg modules. The adaptive oscillating module consists of two simple neuron nodes. Each node effects the other and itself for oscillating and all weights on connections are adaptively learned. The learning algorithm is based on the modified BP method. In this algorithm, each module can independently learn with suitable speed for given input data. the algorithm is applied as a suitable method to decide the minimum size of a network which optimally yields a target waveform.
社団法人電子情報通信学会の論文
1994-09-26