Notes
Absolutely mind blowing the new 2 models that Deep-Mind has recently released, Alpha-Proof, a reinforcement-learning system for formal mathematic reasoning, and AlphaGeometry 2, an improved version of geometry-solving system.
What is the all the fuzz about? Well, these 2 tools were able to solve 4 out 6 problems of International Mathematical Olympiad, that would put them equal to the silver medalists.
Should be noticed that the time frame to solve 3 out of the 4 solved was significantly longer than the time given to the human math Wizz, nonetheless, it is a great achievement and, in the future, as Terence Tao mentioned in an Oxford’s talk, will provide great assistance and help mathematicians to explore their knowledge to a maximum that were not achievable in the recent past!
“…We’re excited for a future in which mathematicians work with AI tools to explore hypotheses, try bold new approaches to solving long-standing problems and quickly complete time-consuming elements of proofs — and where AI systems like Gemini become more capable at math and broader reasoning. …”
