Skip to contentSkip to filters

We use cookies for analytics.

Machine learning tools for mathematical discovery

Adam Zsolt Wagner

Google DeepMind

Recording

1 h

Abstract

Adam Zsolt Wagner surveys machine-learning tools for discovering interesting constructions across a range of mathematical problems. After a brief look at simple reinforcement-learning setups and PatternBoost, the talk concentrates on language-model-based systems such as FunSearch and AlphaEvolve. It compares the strengths and weaknesses of these approaches and considers which method best suits which kind of problem. The work is joint with François Charton, Jordan Ellenberg, Bogdan Georgiev, Javier Gómez-Serrano, Terence Tao and Geordie Williamson.

Topics