Chinese student Haojia Shi won the IMO, while Google DeepMind's AI programs AlphaProof and AlphaGeometry 2 achieved an impressive result at the competition, solving 4 out of 6 problems at a silver medalist level.
AlphaProof uses reinforcement learning like AI algorithms for games, penalizing incorrect moves to improve. In math problems, it verifies reasoning steps alongside solutions using proof assistants algorithms.
[
Collection
]
[
|
...
]