ABSTRACT
An artificial neural network consists of a topology and a set of rules that govern the dynamic aspects of the network. This section contains a detailed treatment of the topology of a neural network, that is, the combined structure of its neurons and connections. It starts with the basic concepts including neurons, connections, and layers, followed by symmetry and high-order aspects. Next, fully and partially connected topologies are discussed, which is complemented by an overview of special topologies like modular, composite, and ontogenic ones. The next section discusses aspects of a formal framework, which is an underlying theme that unites this section in which a balance is sought between clarity and mathematical rigor in the hope of providing a useful basis and reference for the other chapters of this handbook. This section proceeds with a discussion on modular topologies and concludes with theoretical considerations for choosing a neural network topology.
