Mathematical discoveries from program search with large language models

• Published in: Nature (London) Vol. 625; no. 7995; pp. 468 - 475
• Main Authors: Romera-Paredes, Bernardino, Barekatain, Mohammadamin, Novikov, Alexander, Balog, Matej, Kumar, M. Pawan, Dupont, Emilien, Ruiz, Francisco J. R., Ellenberg, Jordan S., Wang, Pengming, Fawzi, Omar, Kohli, Pushmeet, Fawzi, Alhussein
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 18.01.2024; Nature Publishing Group
• Subjects: 639/705/1045; 639/705/117; Algorithms; Applications programs; Artificial Intelligence; Combinatorial analysis; Combinatorics; Computer Science; Design specifications; Feedback loops; Function space; Humanities and Social Sciences; Language; Large language models; Mathematical models; multidisciplinary; Natural language (computers); Neural networks; Science; Science (multidisciplinary); Subject specialists; Task complexity
• ISSN: 0028-0836; 1476-4687; 1476-4687
• DOI: 10.1038/s41586-023-06924-6

Large language models (LLMs) have demonstrated tremendous capabilities in solving complex tasks, from quantitative reasoning to understanding natural language. However, LLMs sometimes suffer from confabulations (or hallucinations), which can result in them making plausible but incorrect statements 1 , 2 . This hinders the use of current large models in scientific discovery. Here we introduce FunSearch (short for searching in the function space), an evolutionary procedure based on pairing a pretrained LLM with a systematic evaluator. We demonstrate the effectiveness of this approach to surpass the best-known results in important problems, pushing the boundary of existing LLM-based approaches 3 . Applying FunSearch to a central problem in extremal combinatorics—the cap set problem—we discover new constructions of large cap sets going beyond the best-known ones, both in finite dimensional and asymptotic cases. This shows that it is possible to make discoveries for established open problems using LLMs. We showcase the generality of FunSearch by applying it to an algorithmic problem, online bin packing, finding new heuristics that improve on widely used baselines. In contrast to most computer search approaches, FunSearch searches for programs that describe how to solve a problem, rather than what the solution is. Beyond being an effective and scalable strategy, discovered programs tend to be more interpretable than raw solutions, enabling feedback loops between domain experts and FunSearch, and the deployment of such programs in real-world applications. FunSearch makes discoveries in established open problems using large language models by searching for programs describing how to solve a problem, rather than what the solution is.