Neurocomputers: an overview of neural networks in VLSI
Manfred Glesner, Werner Pöchmüller · Chapman & Hall eBooks · 1994
Preface I Introduction 1.1 Why Neural Information Processing? 1.2 Dedicated Neural Network Hardware 1.3 About this Book 1.4 Source of Material 2 Categorisation 2.1 Neural Network Type 2.2 Biological Evidence 2.3 Implementation Technology 2.4 Chip Cascadability 2.5 Mapping of Network on Processing Elements 2.6 Flexibility 2.7 Summary 3 Neural Network Simulation 3.l Neural Network Basic Building Blocks 3.2 Artificial Neural Networks 3.3 Summary 4 Digital Building Blocks 4.1 Introduction 4.2 Summation 4.3 Multiplication 4.4 Nonlinearities 4.5 Storage Elements 4.6 Other Elements 4.7 Summary 5 Analog Building Blocks 5.1 Introduction 5.2 Summation 5.3 Multiplication 5.4 Analog Storage Elements 5.5 Non-linear Elements 5.6 Other Elements. 5.7 Summary 6 Optical Building Blocks 6.1 Introduction 6.2 Summation 6.3 Multiplication 6.4 Nonlinearities 6.5 Weight Storage and other Tasks 6.6 Summary 7 Digital Neurocomputers 7.1 Introduction 7.2 Sequential Computers 7.3 Digital Signal Processor Arrays 7.4 Transputer Networks 7.5 RISC Arrays 7.6 SIMD Arrays and Systolic Arrays 7.7 Slice Architectures 7.8 Weightless Neural Networks 7.9 Wafer Scale Implementations 7.10 Summary 8 Analog Neurocomputers 8.1 Introduction 8.2 Artificial Networks near to Biology 8.3 Cellular Neural Networks 8.4 Early NMOS and CMOS Hopfield Network Implementations 8.5 Network Implementations based on Amorphous Silicon 8.6 CCD and Floating-Gate Implementations 8.7 CMOS Networks with Programmable Weights 8.8 Pulse Stream Networks 8.9 Summary 9 Optical Neurocomputers 9.1 Introduction 9.2 Optoelectronic Implementations 9.3 Optical Implementations