神经网络中的粒子权重计算

Granular Weights Computing in Neural Networks

讨论了将粒子信息存入神经网络的机制,以加速网络训练,提高清晰度,提出了语言权重的GNN及语言算法规则。GNN比标准BP网络需要少训练时间来学习已知问题,同时知识的表示更易于理解。

This article discusses the mechanism which graular information will be stored in neural network,in order to speed up network training and increase clarity,then presents GNN based on linguistic weights and linguistic algorithm rules.