Diminishing the number of nodes in multi-layered neural networks
作者： P. Nocera , R. Quelavoine
关键词: Feedforward neural network 、 Multilayer perceptron 、 Recurrent neural network 、 Echo state network 、 Time delay neural network 、 Artificial intelligence 、 Deep learning 、 Computer science 、 Node (networking) 、 Probabilistic neural network
摘要: We propose in this paper two ways for diminishing the size of a multilayered neural network trained to recognise French vowels. The first deals with hidden layers: study variation outputs each node gives us information on its very discrimination power and then allows reduce network. second involves input nodes: by examination connecting weights between nodes following layer, we can determinate which features are actually relevant our classification problem, eliminate useless ones. Through problem recognising vowel /a/, show that obtain reduced structure still learn. >
sci-hub.se 参考文章(1)
L. Bochereau, P. Bourgine, A generalist-specialist paradigm for multilayer neural networks 1990 IJCNN International Joint Conference on Neural Networks. pp. 87- 91 ,(1990) , 10.1109/IJCNN.1990.137828