The new AI, called AlphaGeometry, was trained on millions of synthesized theorems and proofs, outperforming previous methods and reaching the level of the average International Mathematical Olympiad (IMO) gold medalist.
AlphaGeometry uses a neuro-symbolic approach to theorem proving by combining synthetic data generation and custom symbolic engines.
This method allows the AI to generate approximately 10 million steps of synthetic evidence, including the creation of ancillary points that go beyond purely symbolic inference.
The AI was pre-trained on all the synthetic data generated and then further trained by focusing on generating supporting points during the search for evidence.
Source: Ferra
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
