DeepMind introduced two new artificial intelligence models that can solve complex mathematical problems

Google’s artificial intelligence research unit DeepMindWe can tell you that he solved a hitherto unsolvable mathematical problem. we had transferred. Now the company has announced a new development in this field. The company introduced two new artificial intelligence models that can solve complex mathematical problems.

AlphaProof ve AlphaGeometry 2 These models, called , are considered as an important step in the field of mathematical reasoning and problem solving, where existing artificial intelligence systems have difficulties.

The model, called AlphaProof, can prove mathematical expressions in a formal language. This model was developed using a technology called AlphaZero; He had proven himself in games such as chess, shogi and Go. This new model won the silver medal at this year’s International Mathematics Olympiad by succeeding in solving four mathematical problems.

The other model, AlphaGeometry 2, stands out as a model that has improved its ability to solve geometry problems. Trained with much more data than its previous version, AlphaGeometry 2 managed to solve 83 percent of the IMO geometry problems of the past 25 years. This model can solve geometry problems faster and more effectively.

DeepMind states that these new models will advance efforts to develop artificial general intelligence, or AGI. AlphaProof solves mathematical problems using a formal language, while AlphaGeometry 2 can quickly solve complex geometry problems. These advances could help achieve artificial general intelligence (AGI) goals by greatly increasing mathematical thinking abilities.