Neural Computing Architectures

The collection presents neural computing from several angles. Its 18 contributions cover European connectionist research, both foundational and recent, and sit alongside the American PDP work without much overlap in the research they report.

The book is arranged in four parts. The first takes the classical view of neural computing. Another takes a mathematical perspective, drawing on logic, automata theory and probability, and includes the neuron modelled as a logic truth function that could be built directly as a silicon read-only memory. Several chapters focus on cognitive modelling, but most of the work concerns abstract network theory and engineering applications.

Igor Aleksander, the editor, works as a computer science professor at Imperial College London, and he introduces each part. The catalogue gives the subtitle The Design of Brain-Like Machines. Edition note: published by The MIT Press in 1989, first edition, ISBN-13 9780262011105.