Despite the success, there are limitations to the problems FunSearch can handle, as solutions need to be verifiable automatically, ruling out many questions in fields like biology. However, DeepMind's researchers believe this technology will be transformational in computer science and algorithmic discovery. The use of LLMs like FunSearch and AlphaTensor, which found ways to speed up calculations and make key algorithms run faster, is seen as a new age for LLMs, assisting in pushing the boundaries of what is possible in algorithms.
Key takeaways:
- DeepMind's AI division has used an LLM-powered tool called FunSearch to solve two famous unsolved math problems, marking the first time a large language model has been used to solve a long-standing human puzzle.
- FunSearch, short for “searching the function space,” uses an LLM called Codey to write solutions to maths problems using computer programs. The best programs are then combined and fed back to the LLM to improve on, allowing the system to steadily change programs into more powerful ones.
- Large language models (LLMs) are the AI models behind gen AI tools like OpenAI’s ChatGPT and Google’s Bard. They’re not known for making discoveries or providing new facts as they recycle information from their training data to generate responses rather than curating new information.
- Despite the success of FunSearch, the problems need to have solutions that can be verified automatically, which rules out many questions in biology, where hypotheses often need to be tested with lab experiments. However, Deepmind’s researchers are excited about how the technology will impact computer science and algorithmic discovery.