This book approaches VLSI neural networks from a practical viewpoint, using case studies to show the full process of VLSI implementation of a network, and addressing the important issues of learning algorithms and limited precision effects. System aspects and low-power implementation issues are also covered. The authors are all international figures in the field from AT&T Bell Labs, Bellcore and SEDAL.
Overview. Introduction to neural computing. MOS devices and circuits. Analogue VLSI building blocks. Kakadu - a micropower neural network. Supervised learning in an analog framework. A micropower intracardiac electrogram classifier. On-chip perturbation based learning. An analog memory technique. Switched capacitor techniques. A high speed image understanding system. A Boltzmann machine learning system. References. Index.
Overview
This book approaches VLSI neural networks from a practical viewpoint, using case studies to show the full process of VLSI implementation of a network, and addressing the important issues of learning algorithms and limited precision effects. System aspects and low-power implementation issues are also covered. The authors are all international figures in the field from AT&T Bell Labs, Bellcore and SEDAL.