Neural Network Potential

Neural network potentials (NNPs) are machine learning models designed to rapidly and accurately predict the energies and forces within molecular systems, circumventing the computational expense of traditional quantum chemistry methods. Current research emphasizes improving NNP accuracy and efficiency through advanced architectures like graph neural networks and Cartesian tensor networks, incorporating features such as spin and charge states, and developing robust uncertainty quantification techniques, often employing ensemble methods or Bayesian approaches. These advancements are significantly accelerating molecular dynamics simulations across diverse applications, including drug discovery, materials science, and the study of complex chemical reactions.

Papers

November 16, 2024

Constructing accurate machine-learned potentials and performing highly efficient atomistic simulations to predict structural and thermal properties
Junlan Liu, Qian Yin, Mengshu He, Jun Zhou
Full Potential Atomistic Simulation Neural Network Potential Material Separation

September 5, 2024

SpinMultiNet: Neural Network Potential Incorporating Spin Degrees of Freedom with Multi-Task Learning
Koki Ueno, Satoru Ohuchi, Kazuhide Ichikawa, Kei Amii, Kensuke Wakasugi
Multi Task Learning Freedom Matter Neural Network Potential Spin State Spin Prior

June 20, 2024

$\nabla^2$DFT: A Universal Quantum Chemistry Dataset of Drug-Like Molecules and a Benchmark for Neural Network Potentials
Kuzma Khrabrov, Anton Ber, Artem Tsypin, Konstantin Ushenin, Egor Rumiantsev, Alexander Telepov, Dmitry Protasov, Ilya Shenbin, Anton Alekseev, Mikhail Shirokikh, Sergey Nikolenko, Elena Tutubalina, Artur Kadurin
New Benchmark Molecular Property Prediction Drug Like Quantum Chemistry Computer Aided Drug Discovery Neural Network Potential Kohn Sham

May 13, 2024

LATTE: an atomic environment descriptor based on Cartesian tensor contractions
Franco Pellegrini, Stefano de Gironcoli, Emine Küçükbenli
Many Body Interatomic Potential Neural Network Potential Cartesian Tensor

April 18, 2024

Physics-informed active learning for accelerating quantum chemical simulations
Yi-Fan Hou, Lina Zhang, Quanhao Zhang, Fuchun Ge, Pavlo O. Dral
Active Learning Physic Informed Machine Learning Molecular Dynamic Neural Network Potential Quantum Chemical Calculation

March 22, 2024

On the Inclusion of Charge and Spin States in Cartesian Tensor Neural Network Potentials
Guillem Simeon, Antonio Mirarchi, Raul P. Pelaez, Raimondas Galvelis, Gianni De Fabritiis
Curious Price Neural Network Potential Spin State

February 27, 2024

TorchMD-Net 2.0: Fast Neural Network Potentials for Molecular Simulations
Raul P. Pelaez, Guillem Simeon, Raimondas Galvelis, Antonio Mirarchi, Peter Eastman, Stefan Doerr, Philipp Thölke, Thomas E. Markland, Gianni De Fabritiis
Molecular Dynamic Force Field Molecular Simulation Neural Network Potential

November 9, 2023

Data Distillation for Neural Network Potentials toward Foundational Dataset
Gang Seob Jung, Sangkeun Lee, Jong Youl Choi
Crystal Structure Atomistic Simulation Data Distillation Novel Material Standard Datasets Neural Network Potential

September 13, 2023

Crystal structure prediction using neural network potential and age-fitness Pareto genetic algorithm
Sadman Sadeed Omee, Lai Wei, Jianjun Hu
Genetic Algorithm Pareto Optimal Crystal Structure Crystal Structure Prediction Neural Network Potential

May 2, 2023

Single-model uncertainty quantification in neural network potentials does not consistently outperform model ensembles
Aik Rui Tan, Shingo Urata, Samuel Goldman, Johannes C. B. Dietschreit, Rafael Gómez-Bombarelli
Uncertainty Quantification Interatomic Potential Atomistic Simulation Model Ensemble Neural Network Potential Based NIDS

February 28, 2023

CHGNet: Pretrained universal neural network potential for charge-informed atomistic modeling
Bowen Deng, Peichen Zhong, KyuJung Jun, Janosh Riebesell, Kevin Han, Christopher J. Bartel, Gerbrand Ceder
Force Field Interatomic Potential Atomistic Simulation Neural Network Potential Crystal Graph

December 15, 2022

Scalable Bayesian Uncertainty Quantification for Neural Network Potentials: Promise and Pitfalls
Stephan Thaler, Gregor Doehner, Julija Zavadlav
Uncertainty Quantification Common Pitfall Molecular Dynamic Coarse Grained Promise Neural Network Potential

October 14, 2022

Injecting Domain Knowledge from Empirical Interatomic Potentials to Neural Networks for Predicting Material Properties
Zeren Shui, Daniel S. Karls, Mingjian Wen, Ilia A. Nikiforov, Ellad B. Tadmor, George Karypis
Neural Network First Principle Interatomic Potential Atomistic Simulation Material Property Neural Network Potential

January 20, 2022

NNP/MM: Accelerating molecular dynamics simulations with machine learning potentials and molecular mechanic
Raimondas Galvelis, Alejandro Varela-Rial, Stefan Doerr, Roberto Fino, Peter Eastman, Thomas E. Markland, John D. Chodera, Gianni De Fabritiis
Molecular Dynamic Machine Learning Potential Neural Network Potential Biomolecular Simulation Molecular Mechanic

Neural Network Potential

Papers

Constructing accurate machine-learned potentials and performing highly efficient atomistic simulations to predict structural and thermal properties

SpinMultiNet: Neural Network Potential Incorporating Spin Degrees of Freedom with Multi-Task Learning

$\nabla^2$DFT: A Universal Quantum Chemistry Dataset of Drug-Like Molecules and a Benchmark for Neural Network Potentials

LATTE: an atomic environment descriptor based on Cartesian tensor contractions

Physics-informed active learning for accelerating quantum chemical simulations

On the Inclusion of Charge and Spin States in Cartesian Tensor Neural Network Potentials

TorchMD-Net 2.0: Fast Neural Network Potentials for Molecular Simulations

Data Distillation for Neural Network Potentials toward Foundational Dataset

Crystal structure prediction using neural network potential and age-fitness Pareto genetic algorithm

Single-model uncertainty quantification in neural network potentials does not consistently outperform model ensembles

CHGNet: Pretrained universal neural network potential for charge-informed atomistic modeling

Scalable Bayesian Uncertainty Quantification for Neural Network Potentials: Promise and Pitfalls

Injecting Domain Knowledge from Empirical Interatomic Potentials to Neural Networks for Predicting Material Properties

NNP/MM: Accelerating molecular dynamics simulations with machine learning potentials and molecular mechanic