Integrating deep learning with the search for new electron-phonon superconductors represents a burgeoning field of research,where the primary challenge lies in the computational intensity of calculating the electron-p...Integrating deep learning with the search for new electron-phonon superconductors represents a burgeoning field of research,where the primary challenge lies in the computational intensity of calculating the electron-phonon spectral function,α2F(ω),the essential ingredient of Midgal-Eliashberg theory of superconductivity.To overcome this challenge,we adopt a two-step approach.First,we computeα2F(ω)for 818 dynamically stable materials.We then train a deep-learning model to predictα2F(ω),using a training strategy tailored for limited data to temper the model’s overfitting,enhancing predictions.Specifically,we train a Bootstrapped Ensemble of Tempered Equivariant graph neural NETworks(BETE-NET),obtaining an MAE of 0.21,45 K,and 43 K for the moments derived fromα2F(ω):λ,\({\omega}_{\log}\),andω2,respectively,yielding an MAE of 2.5 K for the critical temperature,Tc.Further,we incorporate domain knowledge of the site-projected phonon density of states to impose inductive bias into the model’s node attributes and enhance predictions.This methodological innovation decreases the MAE to 0.18,29 K,and 28 K,respectively,yielding an MAE of 2.1 K for Tc.We illustrate the practical application of our model in high-throughput screening for high-Tc materials.The model demonstrates an average precision nearly five times higher than random screening,highlighting the potential of ML in accelerating superconductor discovery.BETE-NET accelerates the search for high-Tc superconductors while setting a precedent for applying ML in materials discovery,particularly when data is limited.展开更多
基金funded by the U.S.National Science Foundation,Division of Materials Research,under Contract No.NSF-DMR-2118718.A.C.H.and R.G.H.acknowledge additional support from the National Science Foundation under award PHY-1549132(Center for Bright Beams)Part of this research was performed while J.B.G.,A.C.H.,and R.G.H.were visiting the Institute for Pure and Applied Mathematics(IPAM),which is supported by the National Science Foundation(Grant No.DMS-1925919)P.M.D was supported by the U.S.Department of Energy,Office of Science,Office of Basic Energy Sciences,under Award Number DE-SC0022311,during the writing and analysis stages of the project.Computational resources were provided by the University of Florida Research Computing Center.
文摘Integrating deep learning with the search for new electron-phonon superconductors represents a burgeoning field of research,where the primary challenge lies in the computational intensity of calculating the electron-phonon spectral function,α2F(ω),the essential ingredient of Midgal-Eliashberg theory of superconductivity.To overcome this challenge,we adopt a two-step approach.First,we computeα2F(ω)for 818 dynamically stable materials.We then train a deep-learning model to predictα2F(ω),using a training strategy tailored for limited data to temper the model’s overfitting,enhancing predictions.Specifically,we train a Bootstrapped Ensemble of Tempered Equivariant graph neural NETworks(BETE-NET),obtaining an MAE of 0.21,45 K,and 43 K for the moments derived fromα2F(ω):λ,\({\omega}_{\log}\),andω2,respectively,yielding an MAE of 2.5 K for the critical temperature,Tc.Further,we incorporate domain knowledge of the site-projected phonon density of states to impose inductive bias into the model’s node attributes and enhance predictions.This methodological innovation decreases the MAE to 0.18,29 K,and 28 K,respectively,yielding an MAE of 2.1 K for Tc.We illustrate the practical application of our model in high-throughput screening for high-Tc materials.The model demonstrates an average precision nearly five times higher than random screening,highlighting the potential of ML in accelerating superconductor discovery.BETE-NET accelerates the search for high-Tc superconductors while setting a precedent for applying ML in materials discovery,particularly when data is limited.
