Neural Networks Books

This book explains the application of Neural Networks into Principal Component Analysis with clear logic and easy examples. The only problem is that I still think this book is too succinct. I'll have to find papers to see the updating rule for some parameters. But as the first step to understand this topic, this book is good enough.

This is a good book, but my immediate need is for a much more introductory book. I'm just through the first chapter and have already encountered ideas and concepts I've never heard before, and my undergraduate major was mathematics! What's the textbook(s) for the prerequisite(s) for the class that would use "Principles..."?

This book seems to be the most recent and comprehensive survey of SOM applications which are described in detail. Really good technical chapters, could have another one or two on financial business. I have never read such a deep chapter on SOM quality measurements. Overall, good complement to Kohonen's bible.

"The principal focus of this book is the multiple-layer feedforward network (MLFN)." This network is the most common and most used type of neural network. I have all of Masters books and this is his best presentation and source code for the MLFN. Chaper 2. of the book is about a complex number version of the MLFN, but it also includes the common real number versions. Other chapters on Data Preparation, Frequency-Domain, Time/Frequency Localization (Gabor Transform, Fourier Transform, Morlet Wavelets) and applications.

The C++ source code is easy to compile, understand and use. Includes simulated annealing, conjugate gradient algorithms and hybrid learning methods.

What a fun book. I was a polical science major, and expected to be in way over my head with matters relating to, well, real science. Luckily I was not. I picked it up after reading through the nifty stuff on the Freeman Lab site at UC Berkeley. I figured I ought to write up a little somethin' since the other reivew seems as if it's trying to scare people off.

That said, it was hard to tell who the intended audience of the book was. It wasn't for people with any sort of bio backgrounds, as he took the time to explain what hemoglobin was. But he also expected a pretty thorough understanding of epimistological models and how synapses function.

So I'm guessing cog-sci types. Neurology people who never took a philosophy class and are willing to think outside the box might get something out of it, as would philosophy kids who neglected their bio. I think most AI people would find some of it old hat, but a fun read none the less.

Traditional neural networks get pretty skewered. But skewering neural networks has become a lot more popular in the last five years since this book was published, I suppose, so hey. And Freeman does it a lot more eloquently than most.

Now, for the rest of us: As I said, I went into this book knowing relitively little. Which, in retrospect, was good. It was an epiphany. Last time I felt that way was reading Weber's "Prodestant Ethic..." when I was a freshman. As if the wool was being lifted from my eyes.

Knowing nothing about the field I can't really judge how revolutionary or archetypical the theories presented are. But if you know nothing about the field and just want some cerebral stimulation to get you excited about the topic, this is the place to be.

This book provides a systematic account of the foundation of Soft Computing, presenting a unified framework from one side and wealth of applications of Soft Computing technology to solve real world problems. This book contains many new results which until now are not published in single book. The book proves useful to researchers and practitioners interested in using Soft Computing technology, and undergraduates, graduate students.

I consider this book to be very timely. It is used as a college textbook as well as reference by several researchers in the field.

This is not your usual everyday story: oh, no! It goes even further than that and its implications are far more outreaching. Interested? Read on...

A very helpful book for scientists who want to use ANN in Engineering.
The right proportionality of theory and practical development!

I found the beginning of the book a bit tedious but, once I made it past the background material and into the core concepts, it got really exciting. This book proposes a technique for dynamically evolving the structure of a nueral net during training. I haven't tried implementing it, but it's a fascinating theory.

When I was 13, I tried to build an artificial neuron in a test tube, storing the 'weight' of the neuron in electroplating of a mechanical pencil graphite rod. But, I burned myself badly with nitric acid and ended up in the emergency room. I unfortunately wasn't allowed to finish the project.

Reading this unique oral history of neural nets, I discovered that Benard Widrow did the same thing in 1959 (minus the burn part - he was 30 at the time and didn't have to sneak the acid of the science store room at school) and invented the 'Memistor' and Memistor Corporation.

I learned a million other details about neural networks and their history in this book and now feel like I know some of the people interviewed here personally.