About.com Rating

Publisher's Site

AI - Artificial Intelligence - has always had a reputation of being one of those ... you know ... out there technologies. Real programmers worked on payroll or general ledger! But the times, they are a changin'! Today, more than ever, AI is actually the heart of some of the leading, and money making, software in existance. That should change anyone's mind about how 'out there" it is.


Not Just For Research Anymore !

Adam Bosworth, former VP of engineering at BEA Systems and now Google's VP of engineering (and the secret weapon behind VB back when he did his turn at Microsoft) had this to say about AI in a keynote address last year:

"Learning Avalon or Swing isn't going to matter. Machine learning and inference and data mining will. For the first time since computers came along, AI is the mainstream."

And it has taken a long time to get here.

I can remember struggling with the AI language Prolog back in the 70's, and trying to convince management that AI was a worthwhile investment. (Struggling with Prolog was easier.) Few commercial applications were actually built. Back then, it was a great idea before its time.

But if you've called for service from a big company recently, you might have noticed that their voice recognition telephone system - Wonder of Wonders! - actually works now! Amazon has risen from the wreckage of the dot com bust and has taken their place with the ...

ahem ... "profitable" software giants largely on the strength of their data mining and intelligent web software. The "secret sauce" behind Google, the most recent technology eruption on the scene, is almost entirely AI.

AI isn't just for research projects anymore.

Right on time, a practicing programmer, Sara Morgan Rea, has written a new book aimed right at the sweet spot of AI. The subtitle says it all, "Agents, Data Mining, Rule-Based Systems, and Speech Processing". In addition to being the right book at the right time, Rea's book has one more great thing going for it: the examples in the book are all in Visual Basic!

Good choice, Sara!

(VB and C# source code is available on the web site.)

And a great book it is! But it's not for beginners. Rea makes it clear right up front that, "This book is not an introduction to programming ..." In addition to knowing your way around VB.NET, you'll need to know how to work with Javascript, XML and T-SQL. And you'll use Microsoft's new Speech Application - SASDK, the Microsoft XML Parser - MSXML, XML for Analysis SDK, ADO for Multidimensional Objects - ADOMD, Background Intelligent Transfer Service ... This book is a tour of some of the latest Microsoft technologies!

The fact that Rea assumes that you already know your way around Microsoft-style programming helps make this relatively slim book compact and fact filled because doesn't waste time on the fundamentals. But it does have clear, step-by-step instructions for the examples in the book.

Clear, Concise, and Straight Forward Writing

Rea's writing style is organized and straight forward. For example, one of the first things in the book is a definition of the four major topics. These definitions describe what you get in the rest of the book:

Agents are computer programs that perform specific tasks for their owners and may be called upon to interact with other agents. They are usually mobile, meaning that they are able to propagate themselves across networks and execute on remote computers. They also may be able to learn and act proactively. Some may be reactive and simply respond to inputs or other agents. Others may be sent out to do specific things or just gather information. Agent applications can be referenced by several different names. Many people refer to them as bots, spiders, or personal assistants.

Data mining involves the extraction of meaningful information from data by identifying patterns within the data. Patterns discovered can be utilized to make predictions about future data. Uncovering trends in the data allows developers to build powerful applications that can anticipate future needs and streamline existing services.

Rule-based systems determine what programming logic to execute by evaluating a series of rules. Rules are usually represented with IF ... THEN logic and are stored in a central repository.

Speech recognition involves the translation of human speech into machine-readable instructions. Speech-based applications also have a voice component in which the application generates natural-sounding speech.

Along the way, Rea also treats us to insightful comments about what AI is, where it came from, and why it works. As just one example, she tells us the main details about the SALT Forum that created an industry standard for speech recognition (Speech Application Language Tags - SALT) in a side note while a speech recognition application is explained.

Publisher's Site