Self-Organising Neural Networks
Self-Organising Neural Networks
复制标题
DOI：
10.1007/978-1-4471-0825-2_4
复制
发表时间：
1999
期刊：
影响因子：
--
通讯作者：
M. Girolami
中科院分区：
文献类型：
--
作者：
M. Girolami
本章重点介绍具有线性和非线性激活函数的自组织神经网络。许多研究人员已经研究了线性网络的性质及其提取或传递有关观测数据统计信息的能力。由于其额外的复杂性，自然发展到非线性网络需要替代分析工具。非线性网络的涌现行为比线性网络丰富得多。本文简要回顾了源分离背景下的线性和非线性自组织网络。
This chapter focuses on self-organising neural networks with linear and non-linear activation functions. Many researchers have studied the properties of linear networks and their ability to extract or transfer information regarding the statistics of the observed data. The natural progression to non-linear networks requires alternative analysis tools due to their additional complexity. The emergent behaviour of non-linear networks is much richer than their linear counterparts. A brief review of linear and non-linear self-organising networks within the context of source separation is presented here.