Yes, sure, but this doesn't mean that this generalizes to open math research problems, which would be the useful real-world application of this. Otherwise this is just playing known games with known rules, granted better/faster than humans.