Skip to content
Marxos edited this page Jul 3, 2022 · 1 revision

The four questions are answered as

  1. A new neuron is made when 1) unique data enters the network (and senses are made for it) and 2) when a link between neurons is made from correlation (2 activations required).
  2. A new link is made when two neurons, at the same level, fire within their decay time. For time organized data, this is a one directional link.
  3. Modifications to these links occur through 1) the AI:feeler (activated from the AI:understander) or 2) more correlating data enters the network.
  4. a super-node is made when
Different sensory inputs are kept separate with their own neural towers, until they get to the cogitating layer where correlations and links can occur. The cogitating layer is defined by where language has labeled things, so in theory could go very far down into the sensory towers of neurons.

If one wants to make a sophisticated, psychological AI like HAL, one adds more of these cogitating layers, such that the lower level cog-layers are "senses" to the higher level ones. For example, a letter neuron from correlating visual data with phoneme data, now allows .

Clone this wiki locally