Rather than have a tutorial of my own, I have compiled an extensive list of resources that I found invaluable online in building my own Neural Network.  Neural Networks are a collection of layers of artificial neurons which are interconnected in various patterns.  Neural Networks are a black box in that they arrange themselves and rely on brute force trial and error.  One idea that I found fascinating was the possibility of time-dependent Neural Networks, which would allow for looping amongst the neuron layers.  Such loops could simulate memory via finite state machines!

There are many methods to calculating the error in a Neural Network.  While most sites rely on the use of a simple two-layer arrangement, Colin Fahey's site allows for multiple layers.  One approach which I found fascinating was on the AI-Junkie's tutorial, involving a Genetic Algorithm to compute the weights and bias changes, rather than the popular back propagation of nearly every other website.

For assistance in building your own Neural Networks, please see my many examples listed at the bottom of the page.  Each site has tutorials which I have read and were absolutely essential in my success at creating my own Neural Networks.

Perhaps one of the most famous neural network examples is training the network to do the Boolean operation XOR.  Below is the Excel graph of my trained Neural Network's output.

Download the program's binary here.  The Excel data is also available here.

• Colin Fahey's Tutorial on Artificial Neural Networks
• AI-Junkie - Neural Network - tutorial using a Genetic Algorithms for weight calculations
• Aveye.com - Neural Networks
• Artificial Intelligence Depot - Neural Networks Warehouse