Pattern separability of an analog threshold neuron model

DOI Web Site Web Site 参考文献6件

この論文をさがす

説明

<jats:title>Abstract</jats:title><jats:p>In the learning of a neural network model, the excitatory and inhibitory synapses and the threshold are modified. There is little discussion, however, as to what function the neural model can realize in the pattern separation when the synapses or the threshold is modifiable. This paper considers the analog‐threshold neuron model, and discusses the pattern separation performance of the neuron when excitatory synapse, inhibitory synapse and threshold are modifiable. As a result of the discussion, it is shown that the pattern‐separating function of the neuron is maximized when the excitatory and inhibitory synapses as well as the threshold are modifiable. Then a learning system is proposed where the excitatory and inhibitory synapses as well as the threshold are simultaneously modifiable. A model with the proposed learning scheme is analyzed. Finally, it is shown by computer simulation that the proposed model operates as the analysis predicts.</jats:p>

収録刊行物

参考文献 (6)*注記

もっと見る

キーワード

詳細情報 詳細情報について

問題の指摘

ページトップへ