Although seemingly disparate, high-energy nuclear physics (HENP) and machine learning (ML) have begun to merge in the last few years, yielding interesting results. It is worthy to raise the profile of utilizing this novel mindset from ML in HENP, to help interested readers see the breadth of activities around this intersection. The aim of this mini-review is to inform the community of the current status and present an overview of the application of ML to HENP. From different aspects and using examples, we examine how scientific questions involving HENP can be answered using ML.

$\alpha$-clustering structure is a significant topic in light nuclei. A Bayesian convolutional neural network (BCNN) is applied to classify initial nonclustered and clustered configurations, namely Woods-Saxon distribution and three-$\alpha$ triangular (four-$\alpha$ tetrahedral) structure for ${}^{12}\mathrm{C}$ $\left({}^{16}\mathrm{O}\right)$, from heavy-ion collision events generated within a multiphase transport (AMPT) model. Azimuthal angle and transverse momentum distributions of charged pions are taken as inputs to train the classifier. On a multiple-event basis, the overall classification accuracy can reach $95\%$ for ${}^{12}\mathrm{C}/{}^{16}\mathrm{O}+{}^{197}\mathrm{Au}$ events at $\sqrt{S_{NN}}=200\mathrm{GeV}$. With proper constructions of samples, the predicted deviations on mixed samples with different proportions of both configurations could be within $5\%$. In addition, setting a simple confidence threshold can further improve the predictions on the mixed dataset. Our results indicate promising and extensive possibilities of application of machine-learning-based techniques to real data and some other problems in physics of heavy-ion collisions.

The influences of $\alpha$-clustering configurations on the giant dipole resonance (GDR) in hot compound systems are discussed under the framework of an extended quantum molecular dynamics (EQMD) model. The computed GDR spectra in the hot nucleus ${}^{28}\mathrm{Si}$ are split into two peaks which are very consistent with the experimental measurement and are remarkably sensitive to the configurations of projectile ${}^{16}\mathrm{O}$, including the linear chain, kite, square, and tetrahedron configurations. Meanwhile, the sensitivity of the GDR lineshapes to the configurations of ${}^{16}\mathrm{O}$ gradually increases with angular momentum. However, for the fusion system induced by ${}^{20}\mathrm{Ne}$, their GDR lineshapes in ${}^{47}\mathrm{V}$ and ${}^{32}\mathrm{S}$ are mostly independent of the $\alpha$-clustering configurations compared with those in the hot compound systems induced by the projectile ${}^{16}\mathrm{O}$, which may be attributed to their much more similar geometric structures between the square pyramid and trigonal bipyramid configurations in ${}^{20}\mathrm{Ne}$. Additionally, we find that the GDR lineshapes in hot compound systems are sensitive to the geometric distributions of the projectile in phase space when it is compared to those with Woods-Saxon distribution and polarization in projectiles. These results demonstrate that the GDR can be taken as an experimental observable to extract information about the configurations of reaction systems.

Direct photon produced from first proton-neutron ($p\text{−}n$) collision during the early stage of heavy ion reaction is a sensitive probe to reflect energy and momentum distribution of nucleons. In this work, we embedded the hard photon production channel in an extended quantum molecular dynamics (EQMD) model, and took the direct photon as a possible probe to improve namely the Fermi motion in the EQMD model. A possible scheme is offered to handle the dynamical wave packet width within incoherent bremsstrahlung process. Direct photons calculated by our modified EQMD were compared with data of ${}^{14}\mathrm{N}+{}^{12}\mathrm{C}$ at beam energies $E/A$ = 20, 30, and 40 MeV, and it is found that the yield, inverse slope, and angular distribution of direct photons could be reasonably reproduced. In addition, asymmetric reaction systems of ${}^4\mathrm{He} +$ C and ${}^4\mathrm{He} +$ Zn at $E/A$ = 53 MeV are also simulated in this work. It is found that the symmetric angular distribution in the nucleon-nucleon ($N\text{−}N$) center-of-mass (c.m.) frame and the velocity of the $\gamma$-emission source can be reasonably obtained from our method although there is some quantitative difference.

The nuclear excitation by electron capture (NEEC) mechanism is considered to be one of the most effective ways to excite nuclear isomers. Despite being proposed over 50 years ago, direct experimental evidence of NEEC is yet elusive. In this study, we propose an experimental scheme to examine the NEEC process using an electron beam ion trap (EBIT). In an EBIT, highly charged ions are bombarded by an electron beam, which can result in excitation of the nuclei in the trap through mechanisms such as NEEC and Coulomb excitation (CE), etc. Our calculations show that the total rates of NEEC production for some nuclei can reach over 0.1/s, which is higher than these from other mechanisms like CE. The proposed scheme may result in the confirmation of the existence of NEEC, and can also be used to study atomic nuclear excitation related processes.