Abstract:Machine learning force fields (MLFFs) are gradually evolving towards enabling molecular dynamics simulations of molecules and materials with ab initio accuracy but at a small fraction of the computational cost. However, several challenges remain to be addressed to enable predictive MLFF simulations of realistic molecules, including: (1) developing efficient descriptors for non-local interatomic interactions, which are essential to capture long-range molecular fluctuations, and (2) reducing the dimensionality of the descriptors in kernel methods (or a number of parameters in neural networks) to enhance the applicability and interpretability of MLFFs. Here we propose an automatized approach to substantially reduce the number of interatomic descriptor features while preserving the accuracy and increasing the efficiency of MLFFs. To simultaneously address the two stated challenges, we illustrate our approach on the example of the global GDML MLFF; however, our methodology can be equally applied to other models. We found that non-local features (atoms separated by as far as 15$~\r{A}$ in studied systems) are crucial to retain the overall accuracy of the MLFF for peptides, DNA base pairs, fatty acids, and supramolecular complexes. Interestingly, the number of required non-local features in the reduced descriptors becomes comparable to the number of local interatomic features (those below 5$~\r{A}$). These results pave the way to constructing global molecular MLFFs whose cost increases linearly, instead of quadratically, with system size.

Molecular Force Fields with Gradient-Domain Machine Learning (GDML): Comparison and Synergies with Classical Force Fields

Molecular Force Fields with Gradient-Domain Machine Learning: Construction and Application to Dynamics of Small Molecules with Coupled Cluster Forces

sGDML: Constructing Accurate and Data Efficient Molecular Force Fields Using Machine Learning

Machine Learning of Accurate Energy-Conserving Molecular Force Fields

Towards Exact Molecular Dynamics Simulations with Machine-Learned Force Fields

Accurate global machine learning force fields for molecules with hundreds of atoms

BIGDML—Towards accurate quantum machine learning force fields for materials

On the design space between molecular mechanics and machine learning force fields

Ensemble learning of coarse-grained molecular dynamics force fields with a kernel approach.

Towards Linearly Scaling and Chemically Accurate Global Machine Learning Force Fields for Large Molecules

Efficient interatomic descriptors for accurate machine learning force fields of extended molecules

Hybrid Classical/Machine-Learning Force Fields for the Accurate Description of Molecular Condensed-Phase Systems

Molecular Simulations with a Pretrained Neural Network and Universal Pairwise Force Fields

Learning Gradient Fields for Molecular Conformation Generation

Machine Learning Directed Optimization of Classical Molecular Modeling Force Fields

EL-MLFFs: Ensemble Learning of Machine Leaning Force Fields

Accurate machine learning force fields via experimental and simulation data fusion

Machine Learning of coarse-grained Molecular Dynamics Force Fields

Liquid to Crystal Si Growth Simulation Using Machine Learning Force Field

Applications and Advances in Machine Learning Force Fields