genetic algorithm - Has anyone tried to compile code into neural network and evolve it? -


do know if has tried compile high level programming languages (java, c#, etc') recurrent neural network , evolve them?

i mean whole process including memory usage stored in graph of neural net, , i'm talking complex programs (thinking natural language processing problems).

when neural net mean directed weighted graphs spreads activation, , nodes functions of inputs (linear, sigmoid , multiplicative keep simple).

furthermore, people mean in genetic programming or there difference?

neural networks not particularly suited evolving programs; strength tends in classification. if has tried, haven't heard (which considering barely touch neural networks not surprise, active in general ai field @ moment).

the main reason why neural networks aren't useful generating programs represent mathematical equation (numeric, rather functional). given numeric input, numeric output. difficult interpret these in context of program more complicated simple arithmetic.

genetic programming traditionally uses lisp, pure functional language, , programs shown tree diagrams (which similar neural network diagrams - source of confusion?). programs evolved exchanging entire branches of tree (a function , parameters) between programs or regenerating entire branch randomly.

there lot of (and lot of bad) references on both of these topics out there - refrain listing them because isn't clear interested in. wikipedia covers each of these techniques, , starting point.


Comments

Popular posts from this blog

linux - Mailx and Gmail nss config dir -

c# - Is it possible to remove an existing registration from Autofac container builder? -

php - Mysql PK and FK char(36) vs int(10) -