CERN in Geneva and AI developer Mistral AI have signed a statement of intent for a long-term strategic partnership on foundation models for scientific discovery. The first goal is a “Large Physics Model” for high-energy physics.
CERN and Mistral AI have signed a statement of intent for a long-term strategic partnership on foundation models for scientific discovery. CERN contributes its expertise in fundamental physics, scientific data and large-scale computing, while Mistral AI brings large-scale model training. The initial focus is high-energy physics, but the work could also inform AI approaches for the wider scientific community.
A foundation model for collider physics
The central goal is a single pretrained model, described as the first foundation model for collider physics: a “Large Physics Model”. It would be trained on billions of simulated proton-proton collisions at the Large Hadron Collider (LHC) and support several tasks, including reconstructing particle collisions, identifying particles, flagging unexpected signals and evaluating future detector designs.
Today, reconstruction relies on a chain of specialized algorithms that progressively convert detector signals into physics objects. Machine learning already improves individual steps of this chain, and a foundation model could learn across all of it within a single model.
The volume of data involved is considerable. CERN has stored about one exabyte (one million terabytes) of LHC experimental data to date, and the High-Luminosity LHC (HL-LHC), expected to start operating around 2030, will increase the number of collisions tenfold.
Shared AI capabilities for physics
Enrica Porcari, CERN Chief Information Officer, says high-energy physics offers “a unique environment in which artificial intelligence can be developed and tested” with strong rigor, and that the partnership aims to create AI capabilities that the wider physics community can share. Johannes Brandstetter, Vice-President of AI for Engineering at Mistral AI, says the company is proud to work with CERN to explore what large-scale AI models can offer to particle physics, with the potential for new scientific discoveries.
Leave a Reply