A neural network is the combination of artificial intelligence and the knowledge of biological neuron-based networks to create a software/hardware tool to create a specific output or function. A good example of this would be the new Geminoid project. They have created an AI robot that simulates human body patterns (blinking, smiling and breathing).
http://www.foxnews.com/scitech/2011/03/07/creepiest-realistic-robot/?test=faces. Another definition of a neural network is to mimic human/animal patterns. This project will allow humans to simulate human activity without having to use a human life to make it. This could be used in things like police work or other situations that require a human life-form to solve an issue. I believe that while programming this AI robot to mimic human activity, they will stumble upon anomalies that might parallel human anomalies. AI projects typically use mathematical computations and algorithms to achieve what a biological neural network achieve in the human body.
There are many more examples of AI in the software realm. Packages like Neuronet use software-based models and require input/data to compute to achieve an optimal result. This software can take decision-making happen in a nano-second rather than in minutes. Artificial Neural Networks (ANN) take inputs and filter them through a trigger that contains thresholds to obtain an optimal result. ANN works just like a BBN (Biological Neural Network) in that certain inputs will create an output based on certain parameters. The big difference is that our BNN is created by God where ANN is created by humans, by God.
