OpenAI said that its AI model solved the Navier–Stokes equations in a few days

This refers to the Navier–Stokes equations, which describe fluid flow and are used, in particular, to forecast the weather and to model how cream swirls in a cup of coffee. Mathematicians have long been trying to understand whether there are scenarios in which these equations stop working correctly.

A $1 million prize has been set for the answer to this question, and the problem itself is part of the list of seven Millennium Prize Problems, identified at the turn of the century as the most important in mathematics. According to OpenAI, the answer is yes, such scenarios exist, but the company does not intend to claim the award.

OpenAI also published a technical paper describing the result, which mathematicians around the world will now study. The company’s statement immediately sparked debate.

Tristan Buckmaster, a mathematics professor at New York University in the United States, said that together with another mathematician he had made significant progress on the same problem in recent weeks, and suggested that the OpenAI project may have relied on their work. He said the company did not directly answer whether its model used data that he and his coauthor had uploaded to a commercial OpenAI service.

OpenAI employee Sébastien Lubeck said the company’s solution was “significantly” different from their work and that they did not use “specific user data” in the process. OpenAI’s statement also says: “Although this is unlikely, we cannot rule out that anonymized data derived from their use of our products helped improve our models”.

Sources

• Material: theins.ru