Has anyone tried to compile code into a neural network and develop it?

Question

Do you know if anyone tried to compile high-level programming languages ​​(java, C #, etc.) into a recurrent neural network and then develop them?

I mean that the whole process, including the use of memory, is stored in the graphics of the neural network, and I'm talking about complex programs (thinking about the problems of processing a natural language).

When I say “neural network”, I mean oriented weighted graphs that propagate activation, and nodes are the functions of their inputs (linear, sigmoid and multiplicative, to simplify it).

Also, what do people mean in genetic programming, or is there a difference?

Answer 1

; , , . - , ( , , , AI ).

, , , (, ). , . , , .

Lisp, , ( - ?). ( ) .

, , ( ) - , , . Wikipedia , .

Answer 2

. , , - , , "" . , , .

, , , "" ( ), .

Google - .

Answer 3

, .

: http://nexginrc.org/nexginrcAdmin/PublicationsFiles/raid09-sadia.pdf

: , .

Answer 4

, , , , - ( ). sel- .

.

, , , ( ).

Answer 5

You could take a look at gene-programming.org , where they claim that they found some of the results of human competition obtained by genetic programming.

I had not heard of self-developing and self-reproducing programs before. They may exist as special research tools, such as gene-programming.org, but nothing in common for general use. And even if they exist, they are very limited to special operations, such as malware detection, as Allen mentioned.