arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.17255cs.LGcond-mat.mtrl-sci
收藏

通过去噪实现晶体性质预测的自监督学习

Self-supervised learning for crystal property prediction via denoising

Johns Hopkins University Applied Physics Laboratory(约翰斯·霍普金斯大学应用物理实验室)

Alexander New, Nam Q. Le, Michael J. Pekala, Christopher D. Stiles

AI总结:

针对晶体材料性质预测中标注数据稀缺的问题,本文提出晶体去噪自监督学习方法CDSSL,通过从扰动结构中恢复有效结构的预训练任务提升图网络等模型的预测性能,并在多种材料、性质和数据集规模上验证其优于无自监督训练的模型。

摘要:

准确预测晶体材料的性质对于靶向发现至关重要,而这种预测越来越多地借助数据驱动模型来完成。然而,对于许多感兴趣的性质而言,已测定特定性质的材料数量远少于已知材料的数量。为克服这一差距,我们提出了一种用于材料性质预测的新型自监督学习(SSL)策略。我们的方法称为晶体去噪自监督学习(CDSSL),通过一个前置任务对预测模型(例如图网络)进行预训练,该任务要求在给定材料结构的扰动版本时恢复有效的材料结构。我们证明,在不同材料类型、性质和数据集规模下,CDSSL 模型均优于未使用 SSL 训练的模型。

英文摘要:

Accurate prediction of the properties of crystalline materials is crucial for targeted discovery, and this prediction is increasingly done with data-driven models. However, for many properties of interest, the number of materials for which a specific property has been determined is much smaller than the number of known materials. To overcome this disparity, we propose a novel self-supervised learning (SSL) strategy for material property prediction. Our approach, crystal denoising self-supervised learning (CDSSL), pretrains predictive models (e.g., graph networks) with a pretext task based on recovering valid material structures when given perturbed versions of these structures. We demonstrate that CDSSL models out-perform models trained without SSL, across material types, properties, and dataset sizes.

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。

↑