Follow
Idea: a theorem verifier with the mindset of a toddler. When it starts it rejects everything, so you have to show it the two theorems it already knows like a hundred times before it'll admit that some things are, in fact, true
@christianp I'd be surprised if someone doing machine learning hasn't thought of something similar.
@christianp
def run_prover():
while True:
print("Why?")
raw_input()
there, digitoddler 2.0 :>
@christianp So, a bayesian approach?
@christianp And hundred examples to convince it that the definitions are indeed motivated.