Intelligence and Computing

Following is only a comment
“Motivation : Assume nature is not perfect (based on our notions), then
the brain has no reason to be the best method of computation of
information or close to it, hence we should not look at the model of the
working of the brain, if it is just the approximation of the model of the
ideal/higher computation.

Effect of spike as post synaptic potential is equivalent to a rectangular pulse, a neuron essentially looks at the sum of all the inputs and compares it to a value, if the value (voting) is higher than a threshold, then that signature of that information is conveyed to other neurons with different levels of same signature in terms of a weighted rectangular pulse and corresponding input information is then ignored (reset), but following incoming information is again checked (again). This basic unit shall be called a neuron. This neurons forms network which have ability to do ‘certain functions’ depending on the ‘weights’ of the network

To do a specific work, specific structures of these networks could be ‘trained’ to do ‘specific’ tasks. Graphs of these networks can be acyclic or cyclic, acyclic would need learning help from intrinsic parameters from rules called like ‘STDP’ and cyclic would need additional help from the neuron which completes the loop of a directional network

These networks can interact with the surroundings/environment once they are
‘programmed’ to be responsive to the stimuli they can perceive and could
‘think’ if their network evolves mainly temporally and also structurally
(formation of new edges ‘synapses’ in graph) with representation of the
‘events’ as the ‘activity’ of the network”