Efficient Generation of Stable Linear Machine-Learning Force Fields with Uncertainty-Aware Active Learning
arXiv:2303.16538 · doi:10.1088/2632-2153/ace418
Abstract
Machine-learning force fields enable an accurate and universal description of the potential energy surface of molecules and materials on the basis of a training set of ab initio data. However, large-scale applications of these methods rest on the possibility to train accurate machine learning models with a small number of ab initio data. In this respect, active-learning strategies, where the training set is self-generated by the model itself, combined with linear machine-learning models are particularly promising. In this work, we explore an active-learning strategy based on linear regression and able to predict the model's uncertainty on predictions for molecular configurations not sampled by the training set, thus providing a straightforward recipe for the extension of the latter. We apply this strategy to the spectral neighbor analysis potential and show that only tens of ab initio simulations of atomic forces are required to generate stable force fields for room-temperature molecular dynamics at or close to chemical accuracy. Moreover, the method does not necessitate any conformational pre-sampling, thus requiring minimal user intervention and parametrization.
References in corpus (31)
- Canonical sampling through velocity-rescaling
- Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
- SchNet - a deep learning architecture for molecules and materials
- DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
- ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
- E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
- Moment Tensor Potentials: a class of systematically improvable interatomic potentials
- Quantum-Chemical Insights from Deep Tensor Neural Networks
- A Spectral Analysis Method for Automated Generation of Quantum-Accurate Interatomic Potentials
- PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments and Partial Charges
- A Universal Graph Deep Learning Interatomic Potential for the Periodic Table
- Machine learning for molecular simulation
- Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
- Less is more: sampling chemical space with active learning
- Machine Learning Unifies the Modelling of Materials and Molecules
- Active learning of linearly parametrized interatomic potentials
- Physics-inspired structural representations for molecules and materials
- Machine Learning Molecular Dynamics for the Simulation of Infrared Spectra
- Active Learning of Uniformly Accurate Inter-atomic Potentials for Materials Simulation
- Accelerating crystal structure prediction by machine-learning interatomic potentials with active learning
- SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effects
- FCHL revisited: faster and more accurate quantum machine learning
- A neural network interatomic potential for the phase change material GeTe
- sGDML: Constructing Accurate and Data Efficient Molecular Force Fields Using Machine Learning
- A deep potential model with long-range electrostatic interactions
- Committee neural network potentials control generalization errors and enable active learning
- Active learning of reactive Bayesian force fields: Application to heterogeneous hydrogen-platinum catalysis dynamics
- Machine learning force fields based on local parametrization of dispersion interactions: Application to the phase diagram of C
- Predicting tensorial molecular properties with equivariant machine-learning models
- Metadynamics sampling in atomic environment space for collecting training data for machine learning potentials
- Hyperactive Learning (HAL) for Data-Driven Interatomic Potentials