Making LHC data fast enough for real-time physics searches

Including LHC exclusion limits in global fits of beyond-standard-model physics has been too slow to be useful. This work trains a machine-learning model on ATLAS data to create a fast mathematical rule that classifies whether parameter space is allowed or ruled out, then feeds it directly into pMSSM global fits. The result: LHC constraints now run in real time alongside other data, speeding up the search for supersymmetric particles by orders of magnitude.