CERN and Mistral Partner to Build a Foundation Model for Collider Physics

Summary

CERN and Mistral have signed a statement of intent to establish a long-term strategic partnership focused on foundation models for scientific discovery. The initial effort targets high-energy physics and aims to create a single pretrained Large Physics Model for collider physics. The model will be trained on billions of simulated proton-proton collisions from CERN’s Large Hadron Collider and is intended to support several tasks, including reconstructing collisions, identifying particles, flagging unexpected signals, and evaluating future detector designs. CERN currently processes detector signals through a chain of specialised algorithms, although machine learning already improves individual stages; the proposed foundation-model approach would learn across that chain and support multiple tasks in one model. CERN contributes expertise in fundamental physics, scientific data, and large-scale computing, while Mistral contributes large-scale model-training capabilities. CERN stores around one exabyte of LHC experimental data, and the number of collisions is expected to increase tenfold with the High-Luminosity LHC, projected to begin operation around 2030. The partners say the scale and precision of this environment could help develop AI methods useful beyond particle physics, while CERN says the resulting capabilities are intended to be shared with the physics community.