Wang Yutao, Deng Wei, Huang Zhaohui, Li Shuixiang
Department of Mechanics and Engineering Science, College of Engineering, Peking University, Beijing 100871, China.
J Chem Phys. 2022 Apr 21;156(15):154504. doi: 10.1063/5.0088056.
Local structure identification is of great importance in many scientific and engineering fields. However, mathematical and supervised learning methods mostly rely on specific descriptors of local structures and can only be applied to particular packing configurations. In this work, we propose an improved unsupervised learning method, which is descriptor-free, for local structure identification in particle packing. The point cloud is used as the input of the improved method, which directly comes from spatial positions of particles and does not rely on specific descriptors. The improved method constructs an autoencoder based on the point cloud network combined with Gaussian mixture models for dimension reduction and clustering. Numerical examples show that the improved method performs well in local structure identification of quasicrystal disk and sphere packings, achieving comparable accuracy with previous methods. For disordered packings, which have been considered having nearly no local structures, the improved method identifies a nontrivial seven-neighbor motif in the maximally dense random packing of disks and finds acentric structural motifs in the random close packing of spheres, which demonstrate the ability on identification of new and unknown local structures. The improved unsupervised learning method would help obtain information from massive simulation and experimental results as well as devising new order parameters for particle packings.
局部结构识别在许多科学和工程领域都具有重要意义。然而，数学方法和监督学习方法大多依赖于局部结构的特定描述符，并且只能应用于特定的堆积构型。在这项工作中，我们提出了一种改进的无监督学习方法，用于颗粒堆积中的局部结构识别，该方法无需描述符。点云用作改进方法的输入，它直接来自颗粒的空间位置，不依赖于特定描述符。改进后的方法基于点云网络结合高斯混合模型构建自动编码器，用于降维和聚类。数值示例表明，改进后的方法在准晶盘和球体堆积的局部结构识别中表现良好，与先前方法具有相当的准确性。对于被认为几乎没有局部结构的无序堆积，改进后的方法在圆盘的最大密度随机堆积中识别出一个非平凡的七邻域 motif，并在球体的随机密堆积中发现了无中心结构 motif，这证明了其识别新的和未知局部结构的能力。改进后的无监督学习方法将有助于从大量模拟和实验结果中获取信息，并为颗粒堆积设计新的序参量。
