Day 1

Neural networks are a collection of decision making nodes that are connected to one another in a network of pathways that can change and be prioritized and even eliminated based on need. Each decision is small, but all of them piled on top of one another and the using pathways of a given moment make it nearly impossible to predict the outcome in a perfect fashion the result of problem posed to a neural network.

By introducing the ability to increase the number of nodes based on need, and designing the decision making algorithms and design of pathways to be extensible, the limits of the neural network and capabilities of the network to complete problems can be increased dramatically. The use of virtual nodes and virtual nodes that can run recursive networks, links that can be madke from parent networks to child networks, the capabilities of the neural networks can become greater than the capabilities of the hardware the software is running on. It is not the speed of decisions every time that is important, it is the answer that comes back.

When neural networks are combined with preexisting training of parts of the network, recursion, access to external controls and data, the limits of the neural network become the limits of the neural network to change the situation around it.

All neural networks should be subject to oversight by people who have the understanding of the signs and consequences of runaway processing. The best example of a run away neural network is all of life as we know it at the present time.