HEIDi: A deep generative framework for ultra-fast, event-by-event heavy-ion simulations

¹ Frankfurt Institute for Advanced Studies, Ruth-Moufang-Str. 1, D-60438 Frankfurt am Main, Germany
² Xidian-FIAS International Joint Research Center, Giersch Science Center, D-60438 Frankfurt am Main, Germany
³ GSI Helmholtzzentrum für Schwerionenforschung GmbH, Planckstr. 1, D-64291 Darmstadt, Germany
⁴ School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen, P.R. China
⁵ Department of Physics, Duke University, Durham, NC 27708-0305, USA
⁶ Institut für Theoretische Physik, Goethe Universität Frankfurt, Max-von-Laue-Str. 1, D-60438 Frankfurt am Main, Germany
⁷ Helmholtz Research Academy Hesse for FAIR (HFHF), GSI Helmholtz Center for Heavy Ion Physics, Campus Frankfurt, Max-von-Laue-Str. 12, 60438 Frankfurt, Germany

^* e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Published online: 17 April 2026

Abstract

HEIDi, a deep learning-based conditional diffusion model for ultra-fast generation of event-by-event heavy-ion collision output is introduced. Trained on UrQMD outputs, HEIDi is shown to generate point clouds of collision output particles, that accurately reproduce distributions of multiplicity and momentum across 26 different hadron species in UrQMD. Compared to UrQMD cascade simulations, HEIDi achieves a speedup factor of 100. These results demonstrate the potential of HEIDi as a versatile AI tool for both theoretical studies and experimental analyses.