Development of machine learning interatomic potentials for titanium binary alloy systems.

原子模拟关键取决于原子间相互作用势的选择。势的作用是精确地描述原子之间的相互作用，使它们的集体行为引起后来在自然界中观察到的现象。基于高级量子理论的势函数提供了精确的描述，但对于大规模问题缺乏所需的效率。另一方面是经验模型，计算成本低，但不够准确。 该项目的目标是开发钛二元合金系统的原子间相互作用势，以实现具有前所未有精度的大规模原子模拟。该项目将利用一些新的机器学习工具，旨在弥合经典和量子方法之间的差距。这种潜力将使我们能够研究物理学家和工程师都感兴趣的钛合金中多晶的形成。

Atomic simulations critically depend on the choice of the interatomic potential. The role of the potential is to accurately describe interactions between atoms such that their collective behaviour give rise to the phenomena later observed in Nature. The potentials based on high level quantum theories provide accurate description but lacks required efficiency for large scale problems. On the other end of the spectrum are empirical models which are computationally cheap but not accurate enough. The goal of this project is to develop an interatomic potential for titanium binary alloy systems to allow large scale atomic simulations with unprecedented precision. The project will utilize a number of novel machine learning tools which aim to bridge the gap between classical and quantum methods. The potential would allow us to study the formation of polycrystals in titanium alloys which are of interest for both physicists and engineers.