☆ 4.8 Article

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

SCIENCE ADVANCES (2023)

Related references

Note: Only part of the references are listed.

Article Physics, Multidisciplinary

Colossal Enhancement of Atomic Force Response in van der Waals Materials Arising from Many-Body Electronic Correlations

Paul Hauseux et al.

Summary: Understanding van der Waals interactions is important for studying complex materials, and a many-body description can greatly enhance atomic force response. This has significant implications for phonon spectra, free energies, interfacial adhesion, and collective dynamics in materials.

PHYSICAL REVIEW LETTERS (2022)

Add to Collection

Article Multidisciplinary Sciences

Self-consistent determination of long-range electrostatics in neural network potentials

Ang Gao et al.

Summary: Machine learning has the potential to revolutionize molecular simulation by developing efficient and accurate models of interatomic interactions. Neural networks can accurately model such interactions at a fraction of the cost of quantum mechanics-based calculations. However, these networks often fail to describe long-range interactions effectively – a critical factor in dielectric screening and chemical reactivity. To address this issue, the authors introduce a self-consistent field neural network approach, which successfully learns and models the long-range response of molecular systems in potentials.

NATURE COMMUNICATIONS (2022)

Add to Collection

Article Multidisciplinary Sciences

E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials

Simon Batzner et al.

Summary: This paper introduces an E(3)-equivariant deep learning method for accelerating molecular dynamics simulations. The method shows state-of-the-art accuracy and remarkable sample efficiency in faithfully describing the dynamics of complex systems. The Neural Equivariant Interatomic Potentials (NequIP) approach employs E(3)-equivariant convolutions to interact with geometric tensors, resulting in a more information-rich and faithful representation of atomic environments. NequIP outperforms existing models with significantly fewer training data, challenging the commonly held belief about the necessity of massive training sets for deep neural networks.

NATURE COMMUNICATIONS (2022)

Add to Collection

Article Multidisciplinary Sciences

BIGDML-Towards accurate quantum machine learning force fields for materials

Huziel E. Sauceda et al.

Summary: The authors present a new approach called BIGDML for constructing potential energy surfaces of materials. This approach achieves accurate simulations of a wide range of materials using just 10-200 structures for training, with energy accuracies better than 1 meV/atom. The BIGDML model utilizes the full symmetry group of a material and does not assume artificial atom types or localization of atomic interactions, demonstrating high data efficiency and state-of-the-art accuracy.

NATURE COMMUNICATIONS (2022)

Add to Collection

Article Chemistry, Physical

Algorithmic Differentiation for Automated Modeling of Machine Learned Force Fields

Niklas Frederik Schmitz et al.

Summary: This article proposes a new method for reconstructing force fields from atomistic simulation data using algorithmic differentiation techniques in the machine learning modeling process. This approach enables the automatic and efficient usage of new descriptors, resulting in improved computational efficiency and simplified inclusion of additional physical knowledge.

JOURNAL OF PHYSICAL CHEMISTRY LETTERS (2022)

Add to Collection

Article Multidisciplinary Sciences

A fourth-generation high-dimensional neural network potential with accurate electrostatics including non-local charge transfer

Tsz Wai Ko et al.

Summary: Machine learning potentials have limitations in accounting for long-range charge transfer, which can be overcome with the introduction of a fourth-generation high-dimensional neural network potential that includes non-local charge population information. This method provides forces, charges, and energies in excellent agreement with DFT data, significantly extending the applicability of modern machine learning potentials.

NATURE COMMUNICATIONS (2021)

Add to Collection

Article Chemistry, Physical

Learning intermolecular forces at liquid-vapor interfaces

Samuel P. Niblett et al.

Summary: By utilizing a perspective informed by contemporary liquid-state theory, researchers explore training an artificial neural network to describe inhomogeneous, disordered systems, finding that local representations of atomic environments are effective for some properties of liquid-vapor interfaces but fail for properties dependent on unbalanced long-ranged interactions. Explicit models of slowly varying long-ranged interactions improve accuracy in predicting interfacial properties.

JOURNAL OF CHEMICAL PHYSICS (2021)

Add to Collection

Review Chemistry, Multidisciplinary

Machine Learning Force Fields

Oliver T. Unke et al.

Summary: The use of machine learning in computational chemistry has led to significant advancements, particularly in the development of machine learning-based force fields to bridge the gap between accuracy and efficiency. The key concept is to learn the statistical relations between chemical structure and potential energy, without preconceived notions of fixed bonds. Challenges remain for the next generation of machine learning force fields.

CHEMICAL REVIEWS (2021)

Add to Collection

Article Computer Science, Hardware & Architecture

Understanding Deep Learning (Still) Requires Rethinking Generalization

Chiyuan Zhang et al.

Summary: Despite traditional explanations falling short in justifying the excellent generalization of large neural networks, experiments show that state-of-the-art convolutional networks can easily adapt to random labeling during training, indicating a different mechanism contributing to their strong performance in practice.

COMMUNICATIONS OF THE ACM (2021)

Add to Collection

Article Multidisciplinary Sciences

SpookyNet: Learning force fields with electronic degrees of freedom and nonlocal effects

Oliver T. Unke et al.

Summary: SpookyNet is a deep neural network that addresses the issue of electronic degrees of freedom and nonlocality typically ignored in machine-learned force fields. By incorporating chemically meaningful inductive biases and analytical corrections into the network architecture, SpookyNet can improve performance in quantum chemistry and fill important gaps in machine learning models.

NATURE COMMUNICATIONS (2021)

Add to Collection

Article Multidisciplinary Sciences

QM7-X, a comprehensive dataset of quantum-mechanical properties spanning the chemical space of small organic molecules

Johannes Hoja et al.

Summary: The QM7-X dataset contains approximately 4.2 million equilibrium and non-equilibrium structures of small organic molecules, with a comprehensive coverage of various physicochemical properties. It is expected to play a critical role in the development of next-generation machine-learning models for exploring broader regions of chemical compound space and designing molecules with targeted properties.

SCIENTIFIC DATA (2021)

Add to Collection

Review Chemistry, Multidisciplinary

Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems

John A. Keith et al.

Summary: The article discusses the potential impact of machine learning models on chemical sciences and emphasizes the importance of collaboration between expertise in computer science and physical sciences. It provides concise tutorials of computational chemistry and machine learning methods, and demonstrates how they can be used together to provide insightful predictions.

CHEMICAL REVIEWS (2021)

Add to Collection

Article Chemistry, Physical

FCHL revisited: Faster and more accurate quantum machine learning

Anders S. Christensen et al.

JOURNAL OF CHEMICAL PHYSICS (2020)

Add to Collection

Article Multidisciplinary Sciences

From quantum to continuum mechanics in the delamination of atomically-thin layers from substrates

Paul Hauseux et al.

NATURE COMMUNICATIONS (2020)

Add to Collection

Article Chemistry, Physical

A deep neural network for molecular wave functions in quasi-atomic minimal basis representation

M. Gastegger et al.

JOURNAL OF CHEMICAL PHYSICS (2020)

Add to Collection

Review Chemistry, Multidisciplinary

Exploring chemical compound space with quantum-based machine learning

O. Anatole von Lilienfeld et al.

NATURE REVIEWS CHEMISTRY (2020)

Add to Collection

Article Chemistry, Physical

Molecular force fields with gradient-domain machine learning (GDML): Comparison and synergies with classical force fields

Huziel E. Sauceda et al.

JOURNAL OF CHEMICAL PHYSICS (2020)

Add to Collection

Article Chemistry, Multidisciplinary

Quantum machine learning using atom-in-molecule-based fragments selected on the fly

Bing Huang et al.

NATURE CHEMISTRY (2020)

Add to Collection

Article Chemistry, Physical

Molecular force fields with gradient-domain machine learning: Construction and application to dynamics of small molecules with coupled cluster forces

Huziel E. Sauceda et al.

JOURNAL OF CHEMICAL PHYSICS (2019)

Add to Collection

Article Chemistry, Physical

GFN2-xTB-An Accurate and Broadly Parametrized Self-Consistent Tight-Binding Quantum Chemical Method with Multipole Electrostatics and Density-Dependent Dispersion Contributions

Christoph Bannwarth et al.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2019)

Add to Collection

Article Chemistry, Physical

PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments, and Partial Charges

Oliver T. Unke et al.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2019)

Add to Collection

Article Chemistry, Physical

Incorporating long-range physics in atomic-scale machine learning

Andrea Grisafi et al.

JOURNAL OF CHEMICAL PHYSICS (2019)

Add to Collection

Article Multidisciplinary Sciences

Unifying machine learning and quantum chemistry with a deep neural network for molecular wavefunctions

K. T. Schuett et al.

NATURE COMMUNICATIONS (2019)

Add to Collection

Article Computer Science, Interdisciplinary Applications

i-PI 2.0: A universal force engine for advanced molecular simulations

Venkat Kapil et al.

COMPUTER PHYSICS COMMUNICATIONS (2019)

Add to Collection

Article Multidisciplinary Sciences

Quantum mechanics of proteins in explicit water: The role of plasmon-like solute-solvent interactions

Martin Stohr et al.

SCIENCE ADVANCES (2019)

Add to Collection

Article Chemistry, Physical

Non-covalent interactions across organic and biological subsets of chemical space: Physics-based potentials parametrized from machine learning

Tristan Bereau et al.

JOURNAL OF CHEMICAL PHYSICS (2018)

Add to Collection

Article Chemistry, Multidisciplinary

The TensorMol-0.1 model chemistry: a neural network augmented with long-range physics

Kun Yao et al.

CHEMICAL SCIENCE (2018)

Add to Collection

Article Multidisciplinary Sciences

Towards exact molecular dynamics simulations with machine-learned force fields

Stefan Chmiela et al.

NATURE COMMUNICATIONS (2018)

Add to Collection

Article Chemistry, Multidisciplinary

ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost

J. S. Smith et al.

CHEMICAL SCIENCE (2017)

Add to Collection

Article Multidisciplinary Sciences

Quantum-chemical insights from deep tensor neural networks

Kristof T. Schuett et al.

NATURE COMMUNICATIONS (2017)

Add to Collection

Article Multidisciplinary Sciences

Bypassing the Kohn-Sham equations with machine learning

Felix Brockherde et al.

NATURE COMMUNICATIONS (2017)

Add to Collection

Article Multidisciplinary Sciences

Machine learning of accurate energy-conserving molecular force fields

Stefan Chmiela et al.

SCIENCE ADVANCES (2017)

Add to Collection

Article Multidisciplinary Sciences

Wavelike charge density fluctuations and van der Waals interactions at the nanoscale

Alberto Ambrosetti et al.

SCIENCE (2016)

Add to Collection

Proceedings Paper Computer Science, Theory & Methods

Uniform Sampling for Matrix Approximation

Michael B. Cohen et al.

PROCEEDINGS OF THE 6TH INNOVATIONS IN THEORETICAL COMPUTER SCIENCE (ITCS'15) (2015)

Add to Collection

Article Physics, Multidisciplinary

Role of Dispersion Interactions in the Polymorphism and Entropic Stabilization of the Aspirin Crystal

Anthony M. Reilly et al.

PHYSICAL REVIEW LETTERS (2014)

Add to Collection

Article Physics, Multidisciplinary

Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning

Matthias Rupp et al.

PHYSICAL REVIEW LETTERS (2012)

Add to Collection

Article Physics, Multidisciplinary

Accurate and Efficient Method for Many-Body van der Waals Interactions

Alexandre Tkatchenko et al.

PHYSICAL REVIEW LETTERS (2012)

Add to Collection

Review Computer Science, Interdisciplinary Applications

Ab initio molecular simulations with numeric atom-centered orbitals

Volker Blum et al.

COMPUTER PHYSICS COMMUNICATIONS (2009)

Add to Collection

Article Mathematics, Applied

RELATIVE-ERROR CU R MATRIX DECOMPOSITIONS

Petros Drineas et al.

SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS (2008)

Add to Collection

Article Chemistry, Physical

An experimental value for the B1u C-H stretch mode in benzene

U Erlekam et al.

JOURNAL OF CHEMICAL PHYSICS (2006)

Add to Collection

Article Chemistry, Physical

Vibrational assignment of all 46 fundamentals of C60 and C606-:: Scaled quantum mechanical results performed in redundant internal coordinates and compared to experiments

CH Choi et al.

JOURNAL OF PHYSICAL CHEMISTRY A (2000)

Add to Collection

© Peeref 2019-2024. All rights reserved.