Solving the QCD effective kinetic theory with neural networks
Open Access
- 14 November 2025
- journal article
- research article
- Published by Springer Nature in The European Physical Journal C
- Vol. 85 (11) , 1304
Abstract
Event-by-event QCD kinetic theory simulations are hindered by the large numerical cost of evaluating the high-dimensional collision integral in the Boltzmann equation. In this work, we show that a neural network can be used to obtain an accurate estimate of the collision integral in a fraction of the time required for the ordinary Monte Carlo evaluation of the integral. We demonstrate that for isotropic and anisotropic distribution functions, the network accurately predicts the time evolution of the distribution function, which we verify by performing traditional evaluations of the collision integral and comparing several moments of the distribution function. This work sets the stage for an event-by-event modeling of the pre-equilibrium initial stages in heavy-ion collisions.Keywords
All Related Versions
Funding Information
- Ministerio de Ciencia y Tecnología (CEX2023-001318-M, PID2023-152762NB-I00)
- Austrian Science Fund (10.55776/P34455)
- Consellerí de Cultura, Educación e Ordenación Universitaria, Xunta de Galicia (Axudas de apoio ´a etapa predoutoral program (Ref)
- European Research Council (ERC-2018-ADG-835105 YoctoLHC)
- Ö Akademie der Wissenschaften (DOC Fellowship project 27203)
This publication has 48 references indexed in Scilit:
- Limiting attractors in heavy-ion collisionsPhysics Letters B, 2024
- Boosting deep neural networks with geometrical prior knowledge: a surveyArtificial Intelligence Review, 2024
- Pre-equilibrium photons from the early stages of heavy-ion collisionsJournal of High Energy Physics, 2024
- RelaxNet: A structure-preserving neural network to approximate the Boltzmann collision operatorJournal of Computational Physics, 2023
- Collective flow in single-hit QCD kinetic theoryJournal of High Energy Physics, 2021
- Using neural networks to accelerate the solution of the Boltzmann equationJournal of Computational Physics, 2021
- Hydrodynamization in systems with detailed transverse profilesPhysics Letters B, 2020
- A high-bias, low-variance introduction to Machine Learning for physicistsPhysics Reports, 2019
- Nearly isentropic flow at sizeable η/sPhysics Letters B, 2018
- Multilayer feedforward networks are universal approximatorsNeural Networks, 1989