ProbabilisticBoostrapping

Probabilistic Boostrapping

How do you solve a chicken and egg problem if there is one in three chance in any given dinosaur shaped thing spawning the true version of that thing?

Defining The Problem

In my interpreter I'm building, one of my functions is less like a traditional program, and more akin to a hybrid AI system: Genetic Reinforced Bayesian model.

But this creates issues when trying to bootstrap a new version of said program, because your teaching the machine to spawn three different versions of the new version of itself.

Example

Here you create three different versions of the DSL you're wanting to make.

./auto_parser 7 :francais "def encrire(input)
  puts input
end

def encrireln(input)
  puts input
end

def mange(a, b)  
  encrire(%Q(Mange un #{a} en #{b}.))
end

mange(ARGV[0], ARGV[1])" :japanais "def kaku(input)
  puts input
end

def kakuln(input)
  puts input
end

def mange(a, b)  
  kaku(%Q(Mange un #{a} en #{b}.))
end

mange(ARGV[0], ARGV[1])" :allemane "def schreiben(input)
  puts input
end

def schreibenln(input)
  puts input
end

def mange(a, b)  
  schreiben(%Q(Mange un #{a} en #{b}.))
end

mange(ARGV[0], ARGV[1])"