%0 Journal Article
%T A GEOMETRICAL LEARNING ALGORITHM OF BINARY NEURAL NETWORKS FOR CLASSIFICATION<br>二进制神经网络分类问题的几何学习算法
%A ZHU Daming
%A MA Shaohan
%A <br>朱大铭
%A 马绍汉
%J 软件学报
%D 1997
%I 
%X Binary to binary mapping for classification plays an important role in the researches on feed-forward-neural-network learning. In this paper, the geometrical method is employed to work out a new algorithm to train binary neural networks for classification. By analysis of every training vertex's geometrical location, the algorithm always produces a neural network of four layers for a certain classification problem. The advantages of this algorithm are: it runs with guaranteed convergence and goes to converge much more quickly than BP and some other algorithms; it can determine the structure of the neural networks by learning so that a precise classification is carried out.In addition, every neuron generated by the algorithm employs a hard-limit activation function with integer synaptic weights, which makes the actual implementation by VLSI technology more facilitated.
%K Neural networks
%K algorithm
%K convergence
%K training
%K geometry<br>神经网络
%K 算法
%K 收敛
%K 训练
%K 几何
%U http://www.alljournals.cn/get_abstract_url.aspx?pcid=5B3AB970F71A803DEACDC0559115BFCF0A068CD97DD29835&cid=8240383F08CE46C8B05036380D75B607&jid=7735F413D429542E610B3D6AC0D5EC59&aid=E2D45A1A0974EF0945862B537885AF35&yid=5370399DC954B911&vid=5D311CA918CA9A03&iid=5D311CA918CA9A03&sid=6837BC93241057EF&eid=E4EC39E73004B593&journal_id=1000-9825&journal_name=软件学报&referenced_num=6&reference_num=12