Machine learning force fields and coarse-grained variables in molecular dynamics: application to materials and biological systems. (2020-04-15T00:00:00.000000Z)

TL;DR

A review of the current understanding of goals, benefits, and limitations of machine learning techniques for computational studies on atomistic systems, focusing on the construction of empirical force fields from ab-initio databases and the determination of reaction coordinates for free energy computation and enhanced sampling.

Abstract

Machine learning encompasses a set of tools and algorithms which are now becoming popular in almost all scientific and technological fields. This is true for molecular dynamics as well, where machine learning offers promises of extracting valuable information from the enormous amounts of data generated by simulation of complex systems. We provide here a review of our current understanding of goals, benefits, and limitations of machine learning techniques for computational studies on atomistic systems, focusing on the construction of empirical force fields from ab-initio databases and the determination of reaction coordinates for free energy computation and enhanced sampling.

Authors

J. Chodera

3 papers

A. Tkatchenko

11 papers

Michele Ceriotti

2 papers

TL;DR

Abstract

Authors

References231 items

Converged quantum simulations of reactive solutes in superfluid helium: The Bochum perspective.

Inexpensive modeling of quantum dynamics using path integral generalized Langevin equation thermostats.

Markov Models of Molecular Kinetics.

Back-mapping based sampling: Coarse grained free energy landscapes as a guideline for atomistic exploration.

Incorporating long-range physics in atomic-scale machine learning.

Boltzmann generators: Sampling equilibrium states of many-body systems with deep learning

Iterative-Learning Strategy for the Development of Application-Specific Atomistic Force Fields

Past–future information bottleneck for sampling molecular reaction coordinate simultaneously with thermodynamics and kinetics

Artificial Intelligence: A Novel Approach for Drug Discovery.

Deflation reveals dynamical structure in nondominant reaction coordinates

A Performance and Cost Assessment of Machine Learning Interatomic Potentials.

Coupling Molecular Dynamics and Deep Learning to Mine Protein Conformational Space

A unified picture of the covalent bond within quantum-accurate force fields: From organic molecules to metallic complexes’ reactivity

On-the-fly machine learning force field generation: Application to melting points

Anncolvar: Approximation of Complex Collective Variables by Artificial Neural Networks for Analysis and Biasing of Molecular Simulations

A fragmented, permutationally invariant polynomial approach for potential energy surfaces of large molecules: Application to N-methyl acetamide.

Neural networks-based variationally enhanced sampling

Allostery in Its Many Disguises: From Theory to Applications.

Nonlinear Discovery of Slow Molecular Modes using Hierarchical Dynamics Encoders

Local and global perspectives on diffusion maps in the analysis of molecular systems

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

How machine learning can assist the interpretation of ab initio molecular dynamics simulations and conceptual understanding of chemistry

Machine Learning of Coarse-Grained Molecular Dynamics Force Fields

Ab initio thermodynamics of liquid and solid water

Equation of State of Fluid Methane from First Principles with Machine Learning Potentials.

Perspective: Identification of collective variables and metastable states of protein dynamics.

The Dynamic Conformational Landscapes of the Protein Methyltransferase SETD8

Galerkin approximation of dynamical quantities using trajectory data.

Free energy, friction, and mass profiles from short molecular dynamics trajectories

Enspara: Modeling molecular ensembles with scalable data structures and parallel computing

Nonlinear machine learning in simulations of soft and biological materials

Improving collective variables: The case of crystallization.

Girsanov reweighting for metadynamics simulations.

Collective variable discovery and enhanced sampling using autoencoders: Innovations in network architecture and error function design.

Atom-density representations for machine learning.

Redefining the Protein Kinase Conformational Space with Machine Learning.

Accelerating high-throughput searches for new alloys with active learning of interatomic potentials

OpenPathSampling: A Python Framework for Path Sampling Simulations. 2. Building and Customizing Path Ensembles and Sample Schemes

Solid Harmonic Wavelet Scattering for Predictions of Molecule Properties

Butane dihedral angle dynamics in water is dominated by internal friction

Machine Learning of Biomolecular Reaction Coordinates.

Note: Variational Encoding of Protein Dynamics Benefits from Maximizing Latent Autocorrelation

Machine Learning Harnesses Molecular Dynamics to Discover New $\mu$ Opioid Chemotypes

An Automatic Adaptive Importance Sampling Algorithm for Molecular Dynamics in Reaction Coordinates

Solving for high-dimensional committor functions using artificial neural networks

Towards exact molecular dynamics simulations with machine-learned force fields

DeePCG: Constructing coarse-grained models via deep neural networks.

Accelerating crystal structure prediction by machine-learning interatomic potentials with active learning

End-to-end differentiable learning of protein structure

Reweighted autoencoded variational Bayes for enhanced sampling (RAVE).

Binding Pathway of Opiates to μ-Opioid Receptors Revealed by Machine Learning

Machine learning and data science in soft materials engineering

Efficient nonparametric n -body force fields from machine learning

Transferable Neural Networks for Enhanced Sampling of Protein Dynamics.

Advanced simulation techniques for the thermodynamic and kinetic characterization of biological systems

Molecular enhanced sampling with autoencoders: On‐the‐fly collective variable discovery and accelerated free energy landscape exploration

Alchemical and structural distribution based representation for universal quantum machine learning.

SchNet - A deep learning architecture for molecules and materials.

Neural Network Based Prediction of Conformational Free Energies - A New Route toward Coarse-Grained Simulation Models.

Variational encoding of complex dynamics.

Strategies for the exploration of free energy landscapes: Unity in diversity and challenges ahead

Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics

Automatic differentiation in PyTorch

VAMPnets for deep learning of molecular kinetics

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

Diffusion maps tailored to arbitrary non-degenerate Itô processes

Reinforcement Learning Based Adaptive Sampling: REAPing Rewards by Exploring Protein Conformational Landscapes.

Collective variables for the study of long-time kinetics from molecular trajectories: theory and methods.

Data-Driven Model Reduction and Transfer Operator Approximation

Point Cloud Discretization of Fokker-Planck Operators for Committor Functions

Girsanov reweighting for path ensembles and Markov state models.

First-Principles Models for van der Waals Interactions in Molecules and Materials: Concepts, Theory, and Applications.

Navigating at Will on the Water Phase Diagram.

Intrinsic map dynamics exploration for uncharted effective free-energy landscapes

OpenMM 7: Rapid development of high performance algorithms for molecular dynamics

Learning molecular energies using localized graph kernels.