Element-wise representations with ECNet for material property prediction and applications in high-entropy alloys
作者机构:State Key Laboratory of Low-Dimensional Quantum PhysicsDepartment of PhysicsTsinghua UniversityBeijing100084China Frontier Science Center for Quantum InformationBeijing100084China Fracture and Reliability Research InstituteSchool of EngineeringTohoku University6-6-01 AramakiaobaAobakuSendai980-8579Japan College of Physics Science and TechnologyYangzhou UniversityYangzhou225002China Collaborative Innovation Center of Quantum MatterBeijing100084China
出 版 物:《npj Computational Materials》 (计算材料学(英文))
年 卷 期:2022年第8卷第1期
页 面:2409-2418页
核心收录:
学科分类:08[工学] 080502[工学-材料学] 0805[工学-材料科学与工程(可授工学、理学学位)]
基 金:This research was supported by the Tohoku-Tsinghua Collaborative Research Funds,the National Natural Science Foundation of China under Grant No.92270104 the Tsinghua University Initiative Scientific Research Program,Grants-in-Aid for Scientific Research on Innovative Areas on High Entropy Alloys through the grant number P18H05454 of JSPS
主 题:alloys prediction entropy
摘 要:When developing deep learning models for accurate property prediction,it is sometimes overlooked that some material physical properties are insensitive to the local atomic ***,we propose the elemental convolution neural networks(ECNet)to obtain more general and global element-wise representations to accurately model material *** shows better prediction in properties like band gaps,refractive index,and elastic moduli of *** explore its application on high-entropy alloys(HEAs),we focus on the FeNiCoCrMn/Pd systems based on the data of DFT *** knowledge from less-principal element alloys can enhance performance in HEAs by transfer learning ***,the element-wise features from the parent model as universal descriptors retain good accuracy at small data *** this framework,we obtain the concentration-dependent formation energy,magnetic moment and local displacement in some sub-ternary and quinary *** results enriched the physics of those high-entropy alloys.