The use of learning techniques and AI systems holds great promise for the identification and discovery of patterns in mathematics. Even if certain kinds of patterns continue to elude modern ML, we hope our Nature paper can inspire other researchers to consider the potential for AI as a useful tool in pure maths. To replicate the results, anybody can access our interactive notebooks. Reflecting on the incredible mind of Ramanujan, George Frederick James Temple wrote, “The great advances in mathematics have not been made by logic but by creative imagination.” Working with mathematicians, we look forward to seeing how AI can further elevate the beauty of human intuition to new levels of creativity.
This work was done by a team including contributions from Alex Davies, Petar Veličković, Lars Buesing, Sam Blackwell, Daniel Zheng, Nenad Tomašev, Richard Tanburn, Peter Battaglia, Charles Blundell, Xavier Glorot, Matt Overlan, Alyssa Pierce, Natalie Lambert, George Holland, Razia Ahamed, Clemens Meyer, Demis Hassabis and Pushmeet Kohli. We would also like to thank Jan Vonk and Jordan Ellenberg for additional mathematical input.