Point-Cloud Generative Models for Fast Calorimeter Simulation Across Particles and Geometries

This paper summarises recent progress in generative point-cloud fast simulation for highly granular calorimeters in collider experiments. CaloClouds3 generates photon electromagnetic showers independently of detector geometry and runs about 120 times faster than Geant4 on average on a single CPU. The resource also covers CaloHadronic, a transformer-based approach to fast hadronic calorimeter simulation.