Neural Network Architectures
Witold Pedrycz · Wiley Encyclopedia of Computer Science and Engineering · 2008
Abstract In this article, we provide a comprehensive and systematic coverage of architectural aspects of neural networks, which includes generic models of neurons and resulting topologies of the networks. Several computing models of neurons along with their associated processing capabilities are investigated. Architectural investigations include studies of commonly encountered topologies of neural networks and their approximation capabilities. The issue of enhanced interpretability of neural networks is addressed by studying logic‐oriented networks, in which underlying computing involves logic neurons. A suite of design guidelines is also presented.