Accurate machine learning force fields via experimental and simulation data fusion

Abstract

Machine Learning (ML)-based force fields are attracting ever-increasing interest due to their capacity to span spatiotemporal scales of classical interatomic potentials at quantum-level accuracy. They can be trained based on high-fidelity simulations or experiments, the former being the common case. However, both approaches are impaired by scarce and erroneous data resulting in models that either do not agree with well-known experimental observations or are under-constrained and only reproduce some properties. Here we leverage both Density Functional Theory (DFT) calculations and experimentally measured mechanical properties and lattice parameters to train an ML potential of titanium. We demonstrate that the fused data learning strategy can concurrently satisfy all target objectives, thus resulting in a molecular model of higher accuracy compared to the models trained with a single data source. The inaccuracies of DFT functionals at target experimental properties were corrected, while the investigated off-target properties remained largely unperturbed. Our approach is applicable to any material and can serve as a general strategy to obtain highly accurate ML potentials.

References in corpus (27)

• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• Statistically optimal analysis of samples from multiple equilibrium states
• PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments and Partial Charges
• A Performance and Cost Assessment of Machine Learning Interatomic Potentials
• Big Data meets Quantum Chemistry Approximations: The $\Delta$-Machine Learning Approach
• Less is more: sampling chemical space with active learning
• Active learning of linearly parametrized interatomic potentials
• Machine learning prediction errors better than DFT accuracy
• Active Learning of Uniformly Accurate Inter-atomic Potentials for Materials Simulation
• SchNet: A continuous-filter convolutional neural network for modeling quantum interactions
• Machine-learning interatomic potentials for materials science
• TorchMD: A deep learning framework for molecular simulations
• Impact of lattice relaxations on phase transitions in a high-entropy alloy studied by machine-learning potentials
• Machine Learning Inter-Atomic Potentials Generation Driven by Active Learning: A Case Study for Amorphous and Liquid Hafnium dioxide
• Fast and Accurate Uncertainty Estimation in Chemical Machine Learning
• Uncertainty estimation for molecular dynamics and sampling
• Single-model uncertainty quantification in neural network potentials does not consistently outperform model ensembles
• Fast Uncertainty Estimates in Deep Learning Interatomic Potentials
• Towards empirical force fields that match experimental observables
• GemNet-OC: Developing Graph Neural Networks for Large and Diverse Molecular Simulation Datasets
• An Experimentally Driven Automated Machine Learned lnter-Atomic Potential for a Refractory Oxide
• Conceptual and practical bases for the high accuracy of machine learning interatomic potential
• Anomalous thermal expansion in $\alpha$-titanium
• Machine learning potentials for multicomponent systems: The Ti-Al binary system
• Towards Linearly Scaling and Chemically Accurate Global Machine Learning Force Fields for Large Molecules
• Learning Pair Potentials using Differentiable Simulations
• Rapid Computation of Thermodynamic Properties Over Multidimensional Nonbonded Parameter Spaces using Adaptive Multistate Reweighting

Cited by in corpus (2)

• chemtrain: Learning Deep Potential Models via Automatic Differentiation and Statistical Physics
• Morphology-Aware Peptide Discovery via Masked Conditional Generative Modeling