Guidance for the Verification and Validation of Neural Networks
Article Type: Book reports From: Kybernetes, Volume 37, Issue 8
Laura L. Pullum, Brian J. Taylor and Majorie A. Darrah,Wiley-Interscience,Holoken, NJ,2007,i-ix pp. and 133 pp.,ISBN 978-0-470-08457-1,Price: current price at wiley.org.com,
Keywords: Artificial intelligence, Cybernetics, Neural networks
Verification and validation (V&V) of software is a subject where any authoritative text will be welcomed by practitioners. This book has the backing of the IEEE Standard for Software Verification which is referenced as IEEE Std 1012-1998 and contains the results of some five years of research and development. The need to compile it came from the US National Aeronautics and Space Administration activities in safety and mission critical research.
As the title suggests, it is designed to help in the V&V of software systems with an emphasis on adaptive ones and with a particular interest in neural network systems.
We are told that it is case study, which is discussed in some detail, is based on neural networks and contains advice on how adaptive systems may be assured. Many practitioners in this field will want this compilation for their own reading as well as a reference guidebook.
