Abstract:Having access to accurate electron densities in chemical systems, especially for dynamical systems involving chemical reactions, ion transport, and other charge transfer processes, is crucial for numerous applications in materials chemistry. Traditional methods for computationally predicting electron density data for such systems include quantum mechanical (QM) techniques, such as density functional theory. However, poor scaling of these QM methods restricts their use to relatively small system sizes and short dynamic time scales. To overcome this limitation, we have developed a deep neural network machine learning formalism, which we call deep charge density prediction (DeepCDP), for predicting charge densities by only using atomic positions for molecules and condensed phase (periodic) systems. Our method uses the weighted smooth overlap of atomic positions to fingerprint environments on a grid-point basis and map it to electron density data generated from QM simulations. We trained models for bulk systems of copper, LiF, and silicon; for a molecular system, water; and for two-dimensional charged and uncharged systems, hydroxyl-functionalized graphane, with and without an added proton. We showed that DeepCDP achieves prediction R2 values greater than 0.99 and mean squared error values on the order of 10−5e2 Å−6 for most systems. DeepCDP scales linearly with system size, is highly parallelizable, and is capable of accurately predicting the excess charge in protonated hydroxyl-functionalized graphane. We demonstrate how DeepCDP can be used to accurately track the location of charges (protons) by computing electron densities at a few selected grid points in the materials, thus significantly reducing the computational cost. We also show that our models can be transferable, allowing prediction of electron densities for systems on which it has not been trained but that contain a subset of atomic species on which it has been trained. Our approach can be used to develop models that span different chemical systems and train them for the study of large-scale charge transport and chemical reactions.

Searching for new heavy fermions with deep learning

Machine Learning Phases of Strongly Correlated Fermions

Predicting New Heavy Fermion Materials within Carbon-Boron Clathrate Structures

Machine learning in the search for new fundamental physics

Machine Learning-Accelerated Discovery of Novel 2D Ferromagnetic Materials with Strong Magnetization

Deep Learning Model for Finding New Superconductors

A deep learning approach to search for superconductors from electronic bands

Atom table convolutional neural networks for an accurate prediction of compounds properties

Deep dive into machine learning density functional theory for materials science and chemistry

Heuristic machinery for thermodynamic studies of SU(N) fermions with neural networks

Fermionic neural network with effective core potential

Machine Learning Electron Density Prediction Using Weighted Smooth Overlap of Atomic Positions

Machine Learning Detection of Correlations in Snapshots of Ultracold Atoms in Optical Lattices

From individual elements to macroscopic materials: in search of new superconductors via machine learning

A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry

Searching High Temperature Superconductors with the assistance of Graph Neural Networks

Neural Network Prediction of Interatomic Interaction in Multielement Substances and High-Entropy Alloys: A Review

Searching High Spin Polarization Ferromagnet in Heusler Alloy Via Machine Learning

Enhancing Magnetocaloric Material Discovery: A Machine Learning Approach Using an Autogenerated Database by Large Language Models