DeepMind, owned by Google, has introduced its AlphaProof and AlphaGeometry 2 artificial intelligence models, which can solve complex mathematical problems that other AI systems cannot. These models are capable of advanced mathematical reasoning due to their unique approach to problem solving. reports Guard.
AlphaProof uses reinforcement learning and is based on the AlphaZero language model, which has already demonstrated its capabilities in gaming situations. An improved version of the existing system, AlphaGeometry 2, is specifically designed to solve geometry problems. Both models were trained on large amounts of synthetic data, allowing them to effectively solve complex problems.
To test their capabilities, the new AI systems were tasked with solving six International Mathematical Olympiad (IMO) problems. AlphaProof successfully solved three problems, while AlphaGeometry 2 solved one geometry problem. Two problems in combinatorics remained unsolved.
Mathematicians Tim Gowers and Joseph Myers evaluated the submitted solutions and gave them points equivalent to a silver medal at the IMO. This is the first time an AI system has produced results comparable to those that won medals at the International Mathematical Olympiad.
Previously artificial intelligence guess Inevitable collapse due to data cannibalization.
What are you thinking?
Source: Gazeta
Jackson Ruhl is a tech and sci-fi expert, who writes for “Social Bites”. He brings his readers the latest news and developments from the world of technology and science fiction.