摘要
Machine learning(ML)has demon-strated significant potential in en-hancing the predictive capabilities of density functional theory methods.In this study,we develop an ML model for correcting B3LYP-D,a density functional approximation that incorporates dispersion correc-tions for non-covalent interactions.This model utilizes semilocal elec-tron density descriptors,and is trained with accurate reference data for both relative and ab-solute energies.Extensive benchmark tests reveal that the ML correction substantially en-hances the generalization ability of the B3LYP-D functional,improving the predictions of at-omization and dissociation energies for complex molecular systems.It retains the accuracy of B3LYP-D in predicting reaction barrier heights and non-covalent interactions while enabling efficient,fully self-consistent field calculations.This work signifies a promising advancement in the development of ML-corrected functionals that surpass the performance of traditional B3LYP-D.
机器学习在提高密度泛函理论方法的预测能力方面展现出巨大潜力.本研究针对包含非共价相互作用色散校正的B3LYP-D泛函,开发了一种机器学习修正模型.该模型使用了半局域电子密度描述符,并基于相对能量和绝对能量的高精度参考数据进行训练.广泛的基准测试表明,该机器学习模型的校正显著提高了B3LYP-D泛函的泛化能力,改进了对复杂分子体系的原子化能和解离能的预测.在保持B3LYP-D对反应能垒高度和非共价相互作用的预测精度的同时,该模型还能够实现高效的全自洽场计算.这项工作标志着机器学习校正泛函开发取得了重要进展,且其性能超越了传统的B3LYP-D密度泛函近似方法.
基金
supported by the National Natural Science Foundation of China(Nos.22393912,22425301,22373091,22173088)
the AI for Science Foundation of Fudan University(No.Fudan X24AI023)
the Strategic Priority Research Program of the Chinese Academy of Sciences(No.XDB0450101).
