The simulation of the ATLAS detector is a critical piece of computing infrastructure for the particle physics community, and a major capital investment on the part of funding agencies. It has been heavily used in every phase of the lifetime of the detector, from conceptual design to publication as well as upgrades. This colloquium offers an overview of the structure of Monte Carlo simulation in the ATLAS experiment, including the accurate description of geometry and materials and the modeling of physics processes with the Geant4 toolkit. Major upgrades to CERN's accelerator complex are now being undertaken to increase the data taking rate by a factor of 7-10, putting stress on computing resources. The advent of artificial intelligence, particularly generative AI, is currently poised to complement particle transport paradigms which are essentially unchanged for decades. I will discuss how these techniques are making their way into the simulation infrastructure ahead of the high luminosity era.


IFIC Colloquia organizers