DeepMind's AI Generates Novel Solution to Decades-Old Math Problem

The large language model, FunSearch, combines with an evaluator to refine solutions and reveal new scientific knowledge.

DeepMind's large language model (LLM), FunSearch, has generated a novel solution to the decades-old 'cap set puzzle', a complex mathematical problem.
FunSearch combines a pre-trained LLM with an automated evaluator, which checks the LLM's outputs for accuracy, reducing the risk of false information.
The LLM and evaluator work in a continuous cycle, refining initial solutions into new knowledge.
FunSearch's solutions are presented as programs, revealing how the solutions are constructed.
The researchers believe FunSearch could inspire further insights in the scientific community, driving a cycle of improvement and discovery.