Customer Reviews

Naturally Intelligent Systems (Bradford Books)

5 star:		(1)
4 star:		(1)
3 star:		(2)
2 star:		(0)
1 star:		(0)

 

 

It is often said that a picture is worth a thousand words. Well, this book chose the thousand-word route in most cases. Neural nets are ideal candidates for illustrations. So, why they decided to use endless descriptions is beyond me. It is like giving directions over the phone when a map would get to the point much faster.

It also needed to explain more conceptually how neural nets actually work, not just how they are arranged. Examples where the net matches one-to-one with an actual image or pattern are easy to follow, but how they recognize different variations of patterns (variety) I never got a good feel for from this book. However, the description of an Adeline node was pretty good.

This book walks through practical examples of various neural network designs and their relationships among eachother.

The book is a very inspiring introduction to artificial neural computing. It explains the intutive motivation behind the design of almost all major artificial neural models. It explains insight in simple english. This book along with a mathematical rigourous book can provide a very good understanding on the modern neural network research.

It jumped from super-trivial to complex too fast for me, becoming a mere catalog of techniques instead of an explainer of them. For example, I got a feel for how basic image recognition can work, but not for say identifying image patterns of different orientation (rotated) or scalings. I think it should have focused more on making sure the reader had a feel for how neural nets work using basic networks first rather than being so eager to try to cover all the variety of nets there are. More step-by-step state diagrams would have been helpful. Though, the explanation of the Adaline wasn't bad.