Google's DeepMind AlphaProof and AlphaGeometry 2 AI models achieved a silver medalist level in the International Mathematical Olympiad by solving four out of six challenging math problems.
The competition, which traditionally targets young mathematicians, is now extended to include testing machine learning and AI systems' ability in complex math problem-solving.
AlphaProof successfully tackled two algebra problems and a number theory problem, proving the answers, while AlphaGeometry 2 solved the geometry problem but failed at the combinatorics problems.
AlphaProof, using the formal language Lean, trains itself to prove mathematical statements, demonstrating versatile abilities shown previously in games like chess, shogi, and Go.
Collection
[
|
...
]