arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2012.02920cond-mat.mtrl-scics.LGphysics.comp-ph
收藏

用于Li-Si系晶体结构搜索的随机弛豫数据集

Dataset of Random Relaxations for Crystal Structure Search of Li-Si System

Stanford University(斯坦福大学);Google Research(谷歌研究院)

Gowoon Cheon, Lusann Yang, Kevin McCloskey, Evan J. Reed, Ekin D. Cubuk

摘要:

晶体结构搜索是材料设计中的一个长期挑战。我们使用密度泛函理论计算,从随机结构中提出了一个超过100,000个潜在电池阳极材料结构弛豫的数据集。我们通过训练图神经网络来说明数据集的使用,以从随机生成的结构中预测结构松弛。我们的模型直接预测除了力之外的应力,这使得他们能够准确地模拟离子位置和晶格矢量的弛豫。我们发现,在分子动力学模拟上训练的模型无法模拟随机结构的弛豫,而在我们的数据上训练导致相同任务的误差减少了两个数量级。我们的模型能够找到一个实验验证的结构的化学计量举行了培训。我们发现,在训练过程中随机扰动原子位置可以提高模型的准确性和域外泛化能力。

英文摘要:

Crystal structure search is a long-standing challenge in materials design. We present a dataset of more than 100,000 structural relaxations of potential battery anode materials from randomized structures using density functional theory calculations. We illustrate the usage of the dataset by training graph neural networks to predict structural relaxations from randomly generated structures. Our models directly predict stresses in addition to forces, which allows them to accurately simulate relaxations of both ionic positions and lattice vectors. We show that models trained on the molecular dynamics simulations fail to simulate relaxations from random structures, while training on our data leads to up to two orders of magnitude decrease in error for the same task. Our model is able to find an experimentally verified structure of a stoichiometry held out from training. We find that randomly perturbing atomic positions during training improves both the accuracy and out of domain generalization of the models.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑