Author: Toby Segaran
Publisher: O'Reilly,2007
Pages: 360
ISBN: 978-0596529321
Aimed at: Python developers and those interested in AI 
Rating: 4
Pros: A readable and practical approach
Cons: Light on theory
Reviewed by: Mike James

Toby Segaran’s book isn’t really about the hot topic of collective intelligence but this really doesn’t matter. The first thing to say is that his code is in Python which isn’t the most obvious choice of computer language for a book on artificial intelligence but this too doesn’t really matter. Collective intelligence is the idea that small simple algorithms can produce something that is more than the sum of the whole if they communicate and can make collective decisions. You’ve seen collective intelligence in action when birds flock and wheel overhead and ants, termites and other insects build incredible structures and solve complex problems. The topics covered in this book are more mainstream than collective intelligence, it's essentially a book on classical AI expressed in Python. It’s about clustering, decision trees, Bayesian filtering and so on. The explanations are down to earth and practical and the Python code is very readable even if you don’t write Python. The approach is practical enough for you to take the code and actually try to use it. Of course if you do then you will discover that real life is never as easy as theory suggests and you will have to put a lot of work in to making it foolproof. A more subtle problem is that the book is actually light on theory which might make fixing the practical problems more difficult. This is an excellent introduction to intelligent methods in programming but it raises as many practical problems as it solves and as a result you are going to need some additional books.