首页 | 本学科首页   官方微博 | 高级检索  
     

Modular—tree:一个自构筑的神经网络结构
引用本文:陈珂 杨立平. Modular—tree:一个自构筑的神经网络结构[J]. 北京大学学报(自然科学版), 1996, 32(1): 110-119
作者姓名:陈珂 杨立平
作者单位:国家视觉听觉及信息处理实验室北京大学信息科学中心,国家视觉听觉及信息处理实验室北京大学信息科学中心,国家视觉听觉及信息处理实验室北京大学信息科学中心,IBM中国实验室 北京,100871,北京,100871,北京,100871,北京,100085
摘    要:提出了一种新颖的具有自构筑能力的神经网络结构,称之为Modular-tree和两个相应的自构筑算法。在此结构中,任何现存的前馈神经网络均可以作为子网。对于一个给定的学习任务,利用提出的生成算法通过对输入空间递归地划分,自动生成一树状的模块神经网络,从而避免了网络结构预置问题。由于使用了“分治”原理,Modular-tree具有良好的性能及快速训练的能力。此结构已用于多个监督学习问题(包括:标准测试

关 键 词:模块神经网络 自构筑 监督学习

Modular-tree:A Self-architecture Neural Network Architecture
CHEN Ke,YU Xiang,CHI Huisheng. Modular-tree:A Self-architecture Neural Network Architecture[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 1996, 32(1): 110-119
Authors:CHEN Ke  YU Xiang  CHI Huisheng
Abstract:Presented a novel self-architecure modular neural network architecture,called modular-tree for supervised learning.In the architecture,any kind of feedforward neural networks can beemployed as componenets and a modular neural network with the tree structure is generated auto-matically with a growing algorithm by partitioning input space recursively to avoid the problem ofpre-determined structure.Due to the principle of divide-and-conquer used in the proposed architec-ture,the modular-tree can yield both a good performance and significantly fast training.The pro-posed architecture has been applied to several supervised learning tasks including both benchmarkand real-world problems and achieved satisfactory results.
Keywords:modular neural networks  self-architecture  supervised learning
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号