Generative AI for full event simulation of high energy heavy ion collisions at RHIC

High-energy heavy-ion collisions at RHIC recreate the extreme conditions that existed in the universe a microsecond after the Big Bang, where hadrons dissolve into a deconfined medium of quarks and gluons known as the quark-gluon plasma (QGP). The sPHENIX detector was designed and optimized to study these high-energy heavy-ion collisions at RHIC. This presentation will discuss artificial intelligence (AI) models developed for whole-event, full-detector simulations of heavy-ion collisions, in which thousands of particles are produced and traverse the detector. The investigation of Denoising Diffusion Probabilistic Models (DDPMs) for event generation and calorimeter response will be presented and compared with other AI approaches. Ongoing plans to develop generative AI models for hydrodynamic simulations of the medium will also be discussed.