Machine learned force fields typically require manual construction of training sets consisting of thousands of first principles calculations,which can result in low training efficiency and unpredictable errors when ap...Machine learned force fields typically require manual construction of training sets consisting of thousands of first principles calculations,which can result in low training efficiency and unpredictable errors when applied to structures not represented in the training set of the model.This severely limits the practical application of these models in systems with dynamics governed by important rare events,such as chemical reactions and diffusion.We present an adaptive Bayesian inference method for automating the training of interpretable,low-dimensional,and multi-element interatomic force fields using structures drawn on the fly from molecular dynamics simulations.Within an active learning framework,the internal uncertainty of a Gaussian process regression model is used to decide whether to accept the model prediction or to perform a first principles calculation to augment the training set of the model.The method is applied to a range of single-and multi-element systems and shown to achieve a favorable balance of accuracy and computational efficiency,while requiring a minimal amount of ab initio training data.We provide a fully opensource implementation of our method,as well as a procedure to map trained models to computationally efficient tabulated force fields.展开更多
基金B.K.acknowledges generous gift funding support from Bosch Research and partial support from the National Science Foundation under Grant No.1808162L.S.was supported by the Integrated Mesoscale Architectures for Sustainable Catalysis(IMASC),an Energy Frontier Research Center funded by the U.S.Department of Energy,Office of Science,Basic Energy Sciences under Award#DE-SC0012573A.M.K.and S.B.acknowledge funding from the MIT-Skoltech Center for Electrochemical Energy Storage.S.B.T.is supported by the Department of Energy Computational Science Graduate Fellowship under grant DE-FG02-97ER25308.
文摘Machine learned force fields typically require manual construction of training sets consisting of thousands of first principles calculations,which can result in low training efficiency and unpredictable errors when applied to structures not represented in the training set of the model.This severely limits the practical application of these models in systems with dynamics governed by important rare events,such as chemical reactions and diffusion.We present an adaptive Bayesian inference method for automating the training of interpretable,low-dimensional,and multi-element interatomic force fields using structures drawn on the fly from molecular dynamics simulations.Within an active learning framework,the internal uncertainty of a Gaussian process regression model is used to decide whether to accept the model prediction or to perform a first principles calculation to augment the training set of the model.The method is applied to a range of single-and multi-element systems and shown to achieve a favorable balance of accuracy and computational efficiency,while requiring a minimal amount of ab initio training data.We provide a fully opensource implementation of our method,as well as a procedure to map trained models to computationally efficient tabulated force fields.
