Google DeepMind: Alphaproof

Digital Illustration for Google DeepMind's "AlphaProof" AI model. Working with the DeepMind team, I aimed to illustrate how AlphaProof proves complex mathematical statements by translating natural language mathematical problems into formal programming languages in order to explore millions of potential solutions, before eventually mapping a route to the correct answer.

In the first image we explored how to show the process as a game of dominoes, where each piece has code snippets instead of dots—and is a potential partner to the previous piece—creating runs of dominoes eventually leading to the solution.

In the second variation we explored how to show the process as a physical journey for the human mathematician, with the logical path being laid as stepping stones through a desert of mathematical possibilities.

Image 3 & 4 are extra pieces created during the exploration stage, included here to shed some light on the creative process .