ZLZhilong Li

Zhilong Li

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Zhilong Li has a total of 17 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Machine-learning the impact parameter in heavy-ion collisions at s N N = 4 and 11 GeV: A cross-check study with UrQMD, AMPT, and JAM

    by Xiaoqing Yue, Guojun Wei, Yongjia Wang, Zhilong Li, Pengcheng Li, Haojie Xu, Xiangrong Zhu, Qingfeng Li, Fuhu Liu, Yasushi Nara

    By generating heavy-ion collision data with the ultrarelativistic quantum molecular dynamics (UrQMD) model, a multiphase transport (AMPT) model, and the JAM model, the impact parameter $b$ in $\mathrm{Au}+\mathrm{Au}$ collisions at $\sqrt{{s}_{\mathrm{NN}}}=4$ and 11 GeV is reconstructed using supervised learning and unsupervised learning in machine learning (ML). In supervised learning, the performance of ML algorithm is cross-checked by using data obtained from these three transport models. It is found that the typical mean absolute error (MAE), which measures the average magnitude of the absolute difference between the true and predicted $b$, is between 0.2--0.4 fm, even when training ML algorithm with data generated from one model but testing with data from others. While the conventional method (i.e., a polynomial fit to multiplicity as a function of $b$) only works for data generated from the same model. In the classification task, the present ML-based method also shows significantly superior results compared to the traditional approach. In unsupervised learning, the $K$-means clustering algorithm is used to partition collision events directly from experimental-style observables, showing that the algorithm autonomously identifies six clusters corresponding to different centrality classes without relying on predefined model-based binning. Our study demonstrates the strong robustness of using an ML algorithm trained on transport-model data for impact-parameter determination, and indicates that this method has the potential to be generalized to handle real experimental data.

  • Study of yrast and yrare low-lying excited states using machine learning approaches*

    by ZhiLong志龙Li李, BingFeng冰锋Lv吕, YongJia永佳Wang王, C. M. Petrache

    The low-lying excitation energies of the states in even-even nuclei are studied using two modern machine learning algorithms: the Light Gradient Boosting Machine (LightGBM) and Sparse Variational Gaussian Process (SVGP). The obtained results demonstrate that both LightGBM and SVGP perform well on the training and validation datasets when informed by a physics-based feature space. A detailed comparison of the results obtained for and states using the Hartree-Fock-Bogoliubov theory extended by the generator coordinate method and mapped onto a five-dimensional collective quadrupole Hamiltonian shows that both ML algorithms outperform this model in terms of accuracy. The extrapolation capabilities of these algorithms were further validated using newly measured 12 data points of and states, which were not included in the training set. In addition, the partial dependence plot method and the Shapley additive explanations method are used as interpretability tools to analyze the relationship between the input features and model predictions. These tools provide in-depth insights into how the input features influence the prediction of low-lying excitation energies and help identify the most important features that drive the prediction, which are valuable for understanding the low-lying excitation energies.


  • Importance of physical information on the prediction of heavy-ion fusion cross sections with machine learning

    by Zhilong Li, Zepeng Gao, Ling Liu, Yongjia Wang, Long Zhu, Qingfeng Li

    In this work, the Light Gradient Boosting Machine (LightGBM), which is a modern decision tree based machine-learning algorithm, is used to study the fusion cross section (CS) of heavy-ion reaction. Several basic quantities (e.g., mass number and proton number of projectile and target) and the CS obtained from phenomenological formula are fed into the LightGBM algorithm to predict the CS. It is found that, on the validation set, the mean absolute error (MAE) which measures the average magnitude of the absolute difference between log10 of the predicted CS and experimental CS is 0.129 by only using the basic quantities as the input, this value is smaller than 0.154 obtained from the empirical coupled channel model. MAE can be further reduced to 0.08 by including an physical-informed input feature. The MAE on the test set (it consists of 280 data points from 18 reaction systems that not included in the training set) is about 0.19 and 0.53 by including and excluding the physical-informed feature, respectively. We further verify the LightGBM predictions by comparing the CS of Ca40,48+Ni78 obtained from the density-constrained time-dependent Hartree-Fock approach. Our study demonstrates the importance of physical information in predicting fusion cross section of heavy-ion reaction with machine learning.