Reset Filters
Branch: 上海图书馆

Cover is for reference only

Analogue imprecision in MLP training /

Hardware inaccuracy and imprecision are important considerations when implementing neural algorithms. This book presents a study of synaptic weight noise as a typical fault model for analogue VLSI realisations of MLP neural networks and examines the implications for learning and network performance....

Full description

Saved in:
Bibliographic Details
Main Authors: Edwards, Peter J. (Peter John) (Author)
Corporate Authors: World Scientific (Firm)
Group Author: Murray, Alan F.
Published: World Scientific Pub. Co.,
Publisher Address: Singapore :
Publication Dates: 1996.
Literature type: eBook
Language: English
Series: Progress in neural processing ; 4
Subjects:
Neural networks (Computer science)
Perceptrons.
Electronic books.
Online Access: http://www.worldscientific.com/worldscibooks/10.1142/3170#t=toc
Summary: Hardware inaccuracy and imprecision are important considerations when implementing neural algorithms. This book presents a study of synaptic weight noise as a typical fault model for analogue VLSI realisations of MLP neural networks and examines the implications for learning and network performance. The aim of the book is to present a study of how including an imprecision model into a learning scheme as a "fault tolerance hint" can aid understanding of accuracy and precision requirements for a particular implementation. In addition the study shows how such a scheme can give rise to significant performance enhancement.
Carrier Form: 1 online resource (xi,178pages) : illustrations.
Bibliography: Includes bibliographical references (pages 165-172) and index.
ISBN: 9789812830012
CLC: TP183
Contents: 1. Introduction. 1.1. Multi-layer perceptrons. 1.2. Stochastic systems. 1.3. Chapter summary -- 2. Neural network performance metrics. 2.1. Introduction. 2.2. Fault tolerance. 2.3. Generalisation. 2.4. Learning trajectory and speed. 2.5. Chapter summary -- 3. Noise in neural implementations. 3.1. Introduction. 3.2. Implementation errors. 3.3. An implementation error model. 3.4. The mathematical model. 3.5. Chapter summary -- 4. Simulation requirements and environment. 4.1. Introduction. 4.2. Simulation requirements. 4.3. Simulation environment. 4.4. Chapter summary -- 5. Fault tolerance. 5.1. Introduction. 5.2. Test environment. 5.3. Results. 5.4. Chapter summary -- 6. Generalisation ability. 6.1. Introduction. 6.2. Test environment. 6.3. Results. 6.4. Chapter summary -- 7. Learning trajectory and speed. 7.1. Introduction. 7.2. Test aims and method. 7.3. Results. 7.4. Chapter summary -- 8. Penalty terms for fault tolerance. 8.1. Introduction. 8.2. Definition of the penalty terms. 8.3. Learning algorithms. 8.4. Practical issues. 8.5. Chapter summary -- 9. Conclusions. 9.1. Introduction. 9.2. Implications for analogue hardware. 9.3. Synaptic noise as an enhancement scheme. 9.4. Learning hints. 9.5. General conclusions.

Related Book Recommendations