Uncertainty-biased molecular dynamics for learning uniformly accurate interatomic potentials  

作　　者：Viktor Zaverkin David Holzmüller Henrik Christiansen Federico Errica Francesco Alesiani Makoto Takamoto Mathias Niepert Johannes Kästner 

机构地区：[1]NEC Laboratories Europe GmbH,Kurfürsten-Anlage 36,69115 Heidelberg,Germany [2]Institute for Theoretical Chemistry,University of Stuttgart,Pfaffenwaldring 55,70569 Stuttgart,Germany [3]Institute for Stochastics and Applications,University of Stuttgart,Pfaffenwaldring 57,70569 Stuttgart,Germany [4]SIERRA,INRIA Paris,2 rue Simone Iff,75012 Paris,France [5]Institute for Artificial Intelligence,University of Stuttgart,Universitätsstraße 32,70569 Stuttgart,Germany

出　　处：《npj Computational Materials》2024年第1期2379-2396,共18页计算材料学(英文)

基　　金：Funded by Deutsche Forschungsgemeinschaft(DFG,German Research Foundation)under Germany’s Excellence Strategy-EXC 2075-390740016。

摘　　要：Efficiently creating a concise but comprehensive data set for training machine-learned interatomic potentials(MLIPs)is an under-explored problem.Active learning,which uses biased or unbiased molecular dynamics(MD)to generate candidate pools,aims to address this objective.Existing biased and unbiased MD-simulation methods,however,are prone to miss either rare events or extrapolative regions—areas of the configurational space where unreliable predictions are made.This work demonstrates that MD,when biased by the MLIP’s energy uncertainty,simultaneously captures extrapolative regions and rare events,which is crucial for developing uniformly accurate MLIPs.Furthermore,exploiting automatic differentiation,we enhance bias-forces-driven MD with the concept of bias stress.

关 键 词：UNIFORMLY dynamics RARE 

分 类 号：O73[理学—晶体学]