Analog computing for AI?

[rupol2000]

Never used unix. I do know Linux operating system the Successor to Unix.

they stole the code from there, it's not really "successors"

 

[rupol2000]

There is no doubt that neural networks are simple quackery. The human brain is organized not on the basis of neurons, but on the basis of modules that form neurons. Worms also have neurons, this does not mean that the worm has "intelligence". But neural networks do not even reproduce the worm's nervous system. the worm's neurons are capable of potentiation, this is not the case in neural networks. They have absolutely nothing to do with the workings of the intellect.

 

[Wuwei]

There is no doubt that neural networks are simple quackery. The human brain is organized not on the basis of neurons, but on the basis of modules that form neurons. Worms also have neurons, this does not mean that the worm has "intelligence". But neural networks do not even reproduce the worm's nervous system. the worm's neurons are capable of potentiation, this is not the case in neural networks. They have absolutely nothing to do with the workings of the intellect.

You obviously have not been exposed to industrial applications of artificial neural networks (ANN). You are confusing ANN with NN. The operative word here is "Artificial". There is absolutely no attempt to duplicate everything that some living creatures is able to do in the same way they do it.

This link is an interesting application. The system trains itself.

AI Behind AlphaGo: Machine Learning and Neural Network – USC Viterbi School of Engineering

Deep learning allows AlphaGo to continually improve its intelligence by playing a large number of games against itself.​

 

[rupol2000]

The system trains itself.

Nothing is "train" there is a lie. Simple caching in this sense is also "learning"

There is absolutely no attempt to duplicate everything that some living creatures is able to do in the same way they do it.

After the AI lab was taken over by left-wing hippies, yes.

 

[Wuwei]

Nothing is "train" there is a lie. Simple caching in this sense is also "learning"

After the AI lab was taken over by left-wing hippies, yes.

You are wrong on both. You sure have a grudge against ANN.

 

[rupol2000]

What is now used in development most widely and fruitfully - OOP, is much closer to AI than neural networks. OOP did not appear immediately, it took a long time to get to it, starting with neural networks, in the AI laboratory. After the neural networks were finally buried, Minsky sketched the Theory of Frames, which influenced the development of Carl Hewitt, and Object-Oriented Programming was born. This is the pinnacle of laboratory developments. Although the Simula developers perverted the idea, even in this form it works.

 

[rupol2000]

Hewitt is the most prolific AI specialists. This is a genius. He is both the father of logic, object-oriented programming and parallel concurrent computing. Most of computer technology exists only because of him.

 

[rupol2000]

Even the functional paradigm is realized thanks to him. The Scheme language that Sussman developed was just a truncated fork of Hewitt's language. They replaced backtracking with continuations