Google DeepMind’s AI systems can now solve complex math problems

1. Google DeepMind has developed two AI systems, AlphaProof and AlphaGeometry 2, which can solve complex math problems involving advanced reasoning. 2. These systems worked together to solve four out of six problems from the International Mathematical Olympiad (IMO), earning a silver medal. 3. This is the first time any AI system has achieved such a high success rate on these kinds of problems. 4. Math problems involving advanced reasoning are difficult for AI systems because they require forming and drawing on abstractions, complex hierarchical planning, and setting subgoals. 5. AlphaProof is a reinforcement-learning-based system that trains itself to prove mathematical statements in the formal programming language Lean. 6. Gemini AI is fine-tuned to automatically translate math problems from natural language into formal statements, creating a large library of formal math problems. 7. Automating this process is a big step forward for the math community, allowing for greater confidence in the correctness of published results and more collaborative work. 8. AlphaGeometry 2 was optimized to tackle geometry problems and was trained on significantly more synthetic data than its predecessor. 9. Two renowned mathematicians checked the systems' submissions and awarded them a total of 28 points out of a maximum of 42, which would earn a human participant a silver medal. 10. Creating AI systems that can solve more challenging mathematics problems could pave the way for exciting human-AI collaborations and help us learn more about how humans tackle math.