This book introduces and describes Netlab ^Loligo, a neural network development environment that enables highly complex, and highly scalable network designs. The Netlab development and runtime environment was born out of a desire to create biologically inspired neural network structures of much greater complexity, and variety than has been traditionally supported in formulaic ANN models. Netlab is not a formulaic, math-centric environment.

The Netlab environment described by this book provides a set of novel approaches and methods for designing and developing neural networks. For example, among the methods introduced, are two new (patent pending) learning methods:   1.) "weight-to-weight learning" which enables short- and long-term memory to be modeled more correctly, and   2.) “influence learning”, which allows unlimited layering, and completely eliminates restrictions on feedback. Taken together, the various methods discussed in this book will enable the construction of neural networks that more closely resemble the capabilities and structures of biological neural networks.

A new neural network description language, called Noodle™, is also introduced and described. It gives experimenters the tools and power needed to conquer much of the complexity and nuance required of biologically inspired neural network designs. The development metaphor of this language is simply that of electronic device development. Specifically, you build devices (called units) and maintain an inventory of different devices. You then connect those devices together as components in more complex designs, which you also maintain as components to be used in the same way.