AlphaProof: when reinforcement learning meets formal mathematicsThomas Hubert · Google DeepMind27 Mar