DeepMind's artificial intelligence has found a new solution to a complex maths problem for the first time in history

DeepMind

DeepMind has used Large Language Model (LLM) technology to find a new solution to the so-called "multiple covers puzzle", one of the most difficult unsolved mathematical problems.

Here's What We Know

This problem involves determining the maximum number of points in the plane that can be connected by lines such that no three points lie on the same line. The solution to the puzzle has been sought for several decades, but only for small sets of points.

DeepMind's model, called FunSearch, has for the first time discovered new designs for large sets, far beyond anything previously known. This scientific discovery was made possible by the combination of Google's PaLM 2 language model and an automatic fact-checking system.

Although LLM did not solve the problem once and for all, it discovered new facts for science, the scientists claim.

According to the developers, this is the first confirmed discovery in mathematics made with the help of artificial intelligence. It could start a new era in the development of AI and maths.

Source: The Next Web