image Artificial Intelligence is a program that learns from experience.  In the Pathfinder game, the program only learns when you run it.  It doesn’t store the state of the gameboard between execution.  For artificial intelligence in the real world most customers would want the Artificial Intelligence to remember between executions. 

In consumer oriented businesses, for instance the production of beer, there is the human touch.  Learning how various equipment responds, the variation in ingredients like hops, as well many other variables creates a distinctive beer flavor.

One of my favorite real life Artificial Intelligences is the Neurosolutions AI that predicts the flavor of beer.  You can find an awesome article about the creation of beer on their site at:, the site is also a great place to investigate an artificial intelligence in the real world.  Or is it?

This type of AI is called a neural network, which attempts to simulate the thought processes of the human brain.  An explanation of what a neural network can be found at:

If you are not a drinker of beer or find drinking repulsive, think of beer as being similar to the way you might create life saving drugs or produce food, or fuel. 

image Sadly, yesterday, I mentioned a number of languages to use for Artificial Intelligence, I forgot one: MS Excel!  the NeuroSolution software is available in Excel, and you can download an evaluation copy for no cost from (link checked Nov. 29, 2010)

In game design, PathFinder is not a neural network, however, if you want to experiment with a neural network, you can download the sample from here:

Finally, what does this have to with XNA 4.0, Windows Phone 7 and Bing Maps?  Neural Networks can be utilized to perform routing using Bing Maps and traffic reports, along with experience.  For instance, from 6:30 AM to 8:45 AM rush hour in Orange County.  You would use the inroad sensors for estimated speeds and accumulate the travel times over the years.

Now, what about XNA 4.0 or the Windows Phone 7?  I often use this code example as a type of neural network, mostly it is a fancy use of the ArcTangent.  Take a look.  It does make an interesting but simple game.  For the phone, I would use the accelerometer.