Update: May 18, 1996
Some believe that intelligence comes when you have a large amount of axioms linked together.
The problem, with this kind of approach, is that you are on a railway track with no escape possible. You have only one choice, one logic, one result.
Moreover, you will fall into logical traps. Remember Alexander the Great facing the Gordian knot. He overcame the logical trap of the situation by switching to another level of logic. Overcoming the trap of trying to untie the knot, he took his sword and cut it.
Why some are clever and others are stupid? Because the first escaped the logic that the second blindly followed.
Escaping the track is the key factor, the path of evolution, the sign of novelty, of added value.
I'm not familiar with neural nets but I have the feeling that the big difference between the von Neumann's logic and neural networks is here: escaping the track.
Now, let's be more practical:
A basic rule would be:
if var1 = A then actionIf you bug me, I will kick you. This is very primitive of course ;-)
A more subtle rule would be:
if var1 = A then var2 = B if var2 = B then actionI believe that var2 in this example is the starting point, the essence, the birth of intelligence. There is a buffer zone between the stimulus and the action. The action has been delayed. Var2 is a modelling of the world.
Let's replace var2 with a word, a natural language word.
Let's link several rules to one word. Let's link words with other words.
Let's put all these relations into one image:
Let's use one very popular word at this time by saying that words are hubs for the mind.
I've talked about escaping the track. You would ask me where do you put that into your model?
The answer is the image. With the image, you can see from above, you can bypass some hubs, you can look at the most interesting ones. This is the domain of the right side of the brain.
Put the axioms into images then analyze the images.
Return to Main Page