Machine learning for molecular simulation

  title={Machine learning for molecular simulation},
  author={Frank No{\'e} and Alexandre Tkatchenko and Klaus-Robert M{\"u}ller and Cecilia Clementi},
  journal={Annual review of physical chemistry},
Machine learning (ML) is transforming all areas of science. The complex and time-consuming calculations in molecular simulations are particularly suitable for an ML revolution and have already been profoundly affected by the application of existing ML methods. Here we review recent ML methods for molecular simulation, with particular focus on (deep) neural networks for the prediction of quantum-mechanical energies and forces, on coarse-grained molecular dynamics, on the extraction of free… 

Figures from this paper

Deep integration of machine learning into computational chemistry and materials science

An overview of how various aspects of atomistic computational modelling are being transformed by the incorporation of ML approaches is provided and how common workflows to predict structure, dynamics, and spectroscopy are affected by ML.

Perspective on integrating machine learning into computational chemistry and materials science.

An overview of how atomistic computational modeling is being transformed by the incorporation of ML approaches is provided and how common workflows to predict structure, dynamics, and spectroscopy are affected by ML is assessed.

Machine Learning Force Fields: Recent Advances and Remaining Challenges.

The general aspects of ML techniques in the context of creating ML force fields are discussed, and common features of ML modeling and quantum-mechanical approximations, so-called global and local ML models, and the physical differences behind these two classes of approaches are described.

Transferring Chemical and Energetic Knowledge Between Molecular Systems with Machine Learning

This study represents a proof of concept that reliable transfer learning models for molecular systems can be designed paving the way to unexplored routes in prediction of structural and energetic properties of biologically relevant systems.

Prediction of transport property via machine learning molecular movements

A simple supervised ML method is presented to predict the transport properties of materials and it is revealed that two types of molecular mechanisms that contribute to low viscosity are revealed.

Accelerated Simulations of Molecular Systems through Learning of Effective Dynamics.

This work presents a novel framework to advance simulation time scales by up to 3 orders of magnitude by learning the effective dynamics (LED) of molecular systems and believes that the proposed framework provides a dramatic increase to simulation capabilities and opens new horizons for the effective modeling of complex molecular systems.

NNP/MM: Fast molecular dynamics simulations with machine learning potentials and molecular mechanics

NNP/MM, an hybrid method integrating neural network potentials (NNPs) and molecular mechanics (MM) is presented, which allows to simulate a part of molecular system with NNP, while the rest is simulated with MM for efficiency.

Unsupervised Learning Methods for Molecular Simulation Data

This Review provides a comprehensive overview of the methods of unsupervised learning that have been most commonly used to investigate simulation data and indicates likely directions for further developments in the field.

Active Machine Learning for Chemical Dynamics Simulations. I. Estimating the Energy Gradient

The theoretical foundations of a novel ML method which trains from a varying set of atomic positions and their energy gradients, called interpolating moving ridge regression (IMRR), and directly predicts the energy gradient of a new set of Atomic positions are outlined.

Strategies for the construction of machine-learning potentials for accurate and efficient atomic-scale simulations

This tutorial review aims at equipping computational chemists and materials scientists with the required background knowledge for ANN potential construction and application, with the intention to accelerate the adoption of the method so that it can facilitate exciting research that would otherwise be challenging with conventional strategies.



Machine Learning for Molecular Dynamics on Long Timescales

  • F. Noé
  • Computer Science
    Machine Learning Meets Quantum Physics
  • 2020
The learning problems in long timescale MD are defined, successful approaches are presented, and some of the unsolved ML problems in this application field are outlined.

Towards exact molecular dynamics simulations with machine-learned force fields

A flexible machine-learning force-field with high-level accuracy for molecular dynamics simulations is developed, for flexible molecules with up to a few dozen atoms and insights into the dynamical behavior of these molecules are provided.

Perspective: Machine learning potentials for atomistic simulations.

  • J. Behler
  • Materials Science
    The Journal of chemical physics
  • 2016
Recent advances in machine learning (ML) now offer an alternative approach for the representation of potential-energy surfaces by fitting large data sets from electronic structure calculations, which are reviewed along with a discussion of their current applicability and limitations.

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

PhysNet is introduced, a DNN architecture designed for predicting energies, forces, and dipole moments of chemical systems, and it is shown that explicitly including electrostatics in energy predictions is crucial for a qualitatively correct description of the asymptotic regions of a potential energy surface (PES).

Machine Learning of Coarse-Grained Molecular Dynamics Force Fields

CGnets, a deep learning approach, that learns coarse-grained free energy functions and can be trained by a force-matching scheme, is introduced, which shows that CGnets can capture all-atom explicit-solvent free energy surfaces with models using only a few coarse- grained beads and no solvent, while classical coarse-Graining methods fail to capture crucial features of the free energy surface.

Artificial Intelligence Assists Discovery of Reaction Coordinates and Mechanisms from Molecular Dynamics Simulations

This work combines advanced sampling schemes with statistical inference, artificial neural networks, and deep learning to discover molecular mechanisms from MD simulations and proposes practical solutions to make the neural networks interpretable, as illustrated in applications to molecular systems.

Transferable Dynamic Molecular Charge Assignment Using Deep Neural Networks.

HIP-NN charge predictions are many orders of magnitude faster than direct DFT calculations, and combined results provide further evidence that ML (specifically HIP-NN) provides a pathway to greatly increase the range of feasible simulations while retaining quantum-level accuracy.

SchNet - A deep learning architecture for molecules and materials.

The deep learning architecture SchNet is presented that is specifically designed to model atomistic systems by making use of continuous-filter convolutional layers and employs SchNet to predict potential-energy surfaces and energy-conserving force fields for molecular dynamics simulations of small molecules.

Assessment and Validation of Machine Learning Methods for Predicting Molecular Atomization Energies.

A number of established machine learning techniques are outlined and the influence of the molecular representation on the methods performance is investigated, finding the best methods achieve prediction errors of 3 kcal/mol for the atomization energies of a wide variety of molecules.

Variational selection of features for molecular kinetics.

This work presents a method to optimize the feature choice directly, without requiring the construction of the final kinetic model, on a canonical set of 12 fast-folding protein simulations and shows that this procedure leads to more efficient model selection.