Freeman and Skapura provide a practical introduction to artificial neural systems (ANS). The authors survey the most common neural-network architectures and show how neural networks can be used to solve actual scientific and engineering problems and describe methodologies for simulating neural-network architectures on traditional digital computing systems.
If you need to learn neural networks, and you do not know anything about them, this is the first book you should read. If you cannot get this book, then Neural Networks and Deep Learning is a free good second choice.
I like this book, though, because it focuses on various aspects which were seen as critical in the early '90s...and how they tried to tackle it by resorting to neuroscience. The only problem I have is the code given is in Pascal.