Neural Network Emulation of Flow in Heavy-Ion Collisions at Intermediate Energies

TL;DR

This paper develops a deep neural network to emulate heavy-ion collision simulations more efficiently and accurately than Gaussian Process models, enabling faster Bayesian analysis of nuclear reaction data.

Contribution

It introduces a DNN-based emulator for the IBUU heavy-ion collision model, outperforming Gaussian Processes in speed and accuracy within a Bayesian framework.

Findings

DNN accurately emulates IBUU predictions with small training data

DNN achieves about ten times higher accuracy than GP

The approach enables faster Bayesian analysis of collision data

Abstract

Applications of new techniques in machine learning are speeding up progress in research in various fields. In this work, we construct and evaluate a deep neural network (DNN) to be used within a Bayesian statistical framework as a faster and more reliable alternative to the Gaussian Process (GP) emulator of an isospin-dependent Boltzmann-Uehling-Uhlenbeck (IBUU) transport model simulator of heavy-ion reactions at intermediate beam energies. We found strong evidence of DNN being able to emulate the IBUU simulator's prediction on the strengths of protons' directed and elliptical flow very efficiently even with small training datasets and with accuracy about ten times higher than the GP. Limitations of our present work and future improvements are also discussed.