Classification of equation of state in relativistic heavy-ion collisions using deep learning

Abstract

Convolutional Neural Nets, which is a powerful method of Deep Learning, is applied to classify equation of state of heavy-ion collision event generated within the UrQMD model. Event-by-event transverse momentum and azimuthal angle distributions of protons are used to train a classifier. An overall accuracy of classification of 98% is reached for Au+Au events at √ s N N = 11 GeV. Performance of classifiers, trained on events at different colliding energies, is investigated. Obtained results indicate extensive possibilities of application of Deep Learning methods to other problems in physics of heavyion collisions.

Applications of deep learning in high energy nuclear physics

“…向实际情形推广 接着我们把上述研究推广到了更 实际的情形：将流体演化后应发生的粒子化(particalization)、强子的级联多重散射以及共振态衰变 等效应通过混合模型考虑进对碰撞的模拟中 [49] [50] 应用类似的研究方法，用 [25] 中提 出用自编码器从重离子实验原初数据提取核液-气 相变的序参数。 Spinodal Maxwell 图 5 旋节和麦克斯韦状态方程分别对应的谱分布 [51] 。 Figure 5 Example images of the normalized transverse density distributions for the Spinodal and Maxwell EoS [51] . [52] 。该问题的物理设计是从一 个线性西格玛模型(Linear Sigma Model, LSM)出 发，得到描述标量 σ 场演化的随机方程。在此非平 衡演化过程中，相变的行为主要由模型中的有效势 控制，连续过渡行为位于小化学势区，而一阶相变 会发生在大重子化学势区域。而之所以选取标量 σ 场的动力学行为做研究，是因为它的长波模式对应 着手征相变发生的序参数，其半经典演化方程描述 了其典型的非平衡态行为(更多背景和细节可参见 文献 [52][53] ) 。该演化方程为 Fluid Dynamics) [56] 数据，均取得了与 AMPT 数 据相近的分类效果。其具体的网络设计，见图 7。 图 7 探测手征磁效应的卷积神经网络 [55] 。 Figure 7 The deep CNN used to detect CME [55] .…”

An equation-of-state-meter for CBM using PointNet

“…It employs multiple layered Artificial Neural Networks to learn higher dimensional correlations in the data. Machine learning and Deep Learning methods have been widely used both in theory [34][35][36][37][38][39][40][41][42][43][44][45][46][47][48] and in experimental high energy physics [49][50][51][52][53][54][55][56][57][58][59][60][61][62][63][64][65][66][67][68]. Previous studies [18,20] on identifying the QCD phase transitions have shown that Convolutional Neural Network (CNN) based models can accurately classify the underlying equation of state from a hydrodynamic evolution using the p tφ spectra of pions (differential transverse and angular distributions in the transverse plane).…”

Probing criticality with deep learning in relativistic heavy-ion collisions

“…Recently deep learning has been used to study the QCD equation of states by classifying phase transition types, using convolution neural network [66][67][68][69] and point cloud network [70,71]. In heavy ion collisions at low energies, auto-encoder with a single latent variable is also used to study the order parameter of the nuclear liquid-gas phase transition [72].…”