Pointed by Denyse, this article by Arlie Coles is an excellent clarifier of AI.
= = = = = START QUOTE:
To find a math map between inputs and targets, neural networks need a lot of examples of input-target pairs. If you give people a puzzle like, “you want to turn two into four, and five into ten, and you’re only allowed to do one multiplication—what number do you need to multiply by?”, they can spot the perfect solution quickly: two! But neural networks have no such analytical power. We make them approximate the solution, which goes much more slowly, because it involves picking a multiplying number, trying it on all the examples’ input numbers, measuring how close the results were to the targets, and tweaking the number accordingly over and over until using it on the inputs sufficiently approaches the targets. It’s like if in math class, instead of solving directly for x, you could only strategically guess x based on what happened when your guess was used.
= = = = = END QUOTE.
Neural networks are NOT feedback mechanisms. They are not analog and they don’t have instant and constant feedback. The simplest example of instant and constant feedback is a drain trap. Every change in water on either side, and every change in atmospheric pressure on either side, is instantly brought back to balance. Zero delay, zero work needed.
The drain trap was turned into a more complex controller in devices like the Willson buoy or the Lukyanov fluid computer.
Real feedback in electrical or mechanical systems is equally instant and constant, including real nervous systems which are electromechanical.
A less mathy explanation of the neural net strikes me now. A neural net is like the process of editing and refining a manuscript, or designing a car, or auditing a ledger. Inputs from several different writers and designers, or inputs from the same designer at different times, start out as relatively random. A good editor or auditor provides NON-INSTANT feedback to the designers, telling them to subtract VERY or add more chrome or switch a debit from one category to another. The editor has a PURPOSE in mind. The book must serve the needs of a specific audience, or the car must please the company’s desired customers. The editor always mentally or physically checks the product to see how well it serves the PURPOSE, but doesn’t know in advance how the final product will look or act. The process can take days or years depending on the complexity.
With a good enough editor, a million monkeys CAN write Shakespeare, and that’s AI.
Polistra's Mill 