Predicting superconducting temperatures with new hierarchical neural network AI model  

作　　者：Shaomeng Xu Pu Chen Mingyang Qin Kui Jin X-D.Xiang 

机构地区：[1]School of Materials Science and Engineering,Harbin Institute of Technology,Harbin 150001,China [2]Department of Materials Science and Engineering,Southern University of Science and Technology,Shenzhen 518055,China [3]Beijing National Laboratory for Condensed Matter Physics,Institute of Physics,Chinese Academy of Sciences,Beijing 100190,China

出　　处：《Frontiers of physics》2025年第1期165-172,共8页物理学前沿(英文版)

基　　金：support from the National Key R&D Program of China(Grant No.2022YFB3807700);the Shenzhen Fundamental Research Funding(Nos.JCYJ20220818100612027 and JCYJ20220818100613028);the Major Science and Technology Infrastructure Project of Shenzhen Material Genome Big-Science Facilities Platform.

摘　　要：Superconducting critical temperature is the most attractive material property due to its impact on the applications of electricity transmission,railway transportation,strong magnetic fields for nuclear fusion and medical imaging,quantum computing,etc.The ability to predict its value is a constant pursuit for condensed matter physicists.We developed a new hierarchical neural network(HNN)AI algorithm to resolve the contradiction between the large number of descriptors and the small number of datasets always faced by neural network AI approaches to materials science.With this new HNN-based AI model,a much-increased number of 909 universal descriptors for inorganic compounds,and a dramatically cleaned database for conventional superconductors,we achieved high prediction accuracy with a test R^(2)score of 95.6%.The newly developed HNN model accurately predicted T_(c)of 45 new high-entropy alloy superconductors with a mean absolute percent error below 6%compared to the experimental data.This demonstrated a significant potential for predicting other properties of inorganic materials.

关 键 词：conventional superconducting critical temperature hierarchical neural network universal descriptors artificial intelligence 

分 类 号：TP1[自动化与计算机技术—控制理论与控制工程]