Here is a collection of books I have myself read and would recommend as informative reading. The list is not exhaustive it is merely a personal appreciation of the moment.

1. "Object Oriented Neural Network in C++" by Joey Rogers

The book is very well done for any neophytes in C++ or Object Oriented. I believe it is a good reference as it is small, easy to understand and fast to read. Also, it does only explain a way of Object Oriented implementation of several neural networks types along with some samples.

On the minus side, the book does not really explain how to use nor the purpose we may found in choosing a specific neural network type from another.

On the code aspect, the object oriented approach described in the book is very much academic (I would personally not consider it very practical).

 

2. "Neural Network Fundamentals with Graphs, Algorithms, and applications"  by N.K. Bose & P.Liang

   When I started this book I really learn something. I am still reading it and I consider it as one of the best book I have encountered (OK, I haven't been reading that much neither) but It is still one of my favorite. This is why I cannot wait to insert the book and this comment.

    

3. "Practical Neural Network Recipes in C++" by Timothy Masters

This book is quite complete on the topic: discussing a way to elude local minima, how to create training sets etc ...

Sometimes quite obscure concerning matters that should really be explained easily. It sometimes need to read several times the same chapter or paragraph in order to fully understand the meaning.

On the code aspect, do not be fooled! what is in C++ is merely an awful unreadable C code encapsulated in a so-called C++ class (it has C++ only the name class and the extension) the rest being some terrible old-fashioned C-coding from a pure mathematician if you can see what I mean. I have tried several time to understand some functions but "NO Way Jose!" the code is totally unreadable.

 

4. "Mind an introduction to cognitive science" by Paul Thagard

This is an excellent introduction to the applied epistemology. It is not intended to software engineers only but anyone who has any interests in knowledge.

I would personally have it more detailed as it gives you the taste but the taste only... therefore you will have to read other more detailed books.

 

5. "Computational philosophy of science" by Paul Thagard

In this book we enter the deep realm of epistemology I really got the impression of entering a new world. This is very well explained and detailed book. One I must believe every academician should have read. We enter in details the notion of abduction and the related information. This is really a must to read in order to taste how knowledge appear. (the book is only available at the MIT press)

 

6. "Conceptual Revolution" by Paul Thagard

This is the cherry on the cake... at least for the programmer. We explore and simulate via neural network tool how revolution in the history of knowledge came and won: Lavoisier's theory of oxygen versus Phlogiston but also the plate tectonics' theory, etc etc...  It read really easily and is very clear for all concepts. (the book is only available at the MIT press)

 

7. "Bayesian Networks and Decision Graphs" by Finn V. Jensen

   Although I am still reading the book, I have already some comments. The reader might perceive the book as easily understandable for a neophyte in bayesian network but it is not! I believe the intention of the author is to make the book as easy as possible but he is then avoiding or skipping explanation of some importance in the benefit of examples.

    

8. "Learning Bayesian Network" by Richard E. Neapolitan

   I am still reading the book, and it is great stuff... I really enjoy reading it. Although some academic background is required. But understanding the book is really worth.

    

9. "The Nonlinear Workbook" by Willi-Hans Steeb

   The book is great for a small and fast reference or reminder of neural network, genetic algorithm, fuzzy logic with a feeble attempt of some C++ coding (style is awful) but helpful nonetheless. Great to have nearby in case of. I really would not consider it as a source of knowledge but a source of shortly condensed information regarding the nonlinear concepts. Like it states it is a workbook and a good one!

    

10. "The Temporal Logic of Reactive and Concurrent system" by Zohar Manna & Amir Pnueli

    This was for many years my night book... I really personally enjoy the concepts and ideas of temporal logic. Sound like the book does not belong out here - well be wrong as a neural network will have to maintain for a while the information or at least  remnant of it. Anyway it is really great book that every informatician would enjoy reading.

     

11. "Stochastic Models of Neural Networks" by Claudio Turchetti

    I really liked to read this book. it is quite interesting and for once we can immediately see a hardware of our neural network - makes it less theoretic. The book requires some time to read, think and process a lot of information in a few lines. It is not always that obvious to apprehend the meaning of the author. I would have welcome a more detailed explanation of many functions and demonstrations. Certainly to recommend for reading and understanding (if time is available). The book makes me believe it is a syllabus of a course given by the author... In the case we really need to follow the course or have some times to understand (not even all of it) and digest the information.

 

 

 

 

 

 

Should you have any book to recommend, please let me know, you can always mail me at c4antares at ainenn dot org