Graph Dynamical Networks for Unsupervised Learning of Atomic Scale Dynamics in Materials

Abstract

Understanding the dynamical processes that govern the performance of functional materials is essential for the design of next generation materials to tackle global energy and environmental challenges. Many of these processes involve the dynamics of individual atoms or small molecules in condensed phases, e.g. lithium ions in electrolytes, water molecules in membranes, molten atoms at interfaces, etc., which are difficult to understand due to the complexity of local environments. In this work, we develop graph dynamical networks, an unsupervised learning approach for understanding atomic scale dynamics in arbitrary phases and environments from molecular dynamics simulations. We show that important dynamical information can be learned for various multi-component amorphous material systems, which is difficult to obtain otherwise. With the large amounts of molecular dynamics data generated everyday in nearly every aspect of materials design, this approach provides a broadly useful, automated tool to understand atomic scale dynamics in material systems.

25 + 7 pages, 5 + 3 figures

References in corpus (19)

• TensorFlow: A system for large-scale machine learning
• Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
• On representing chemical environments
• Deep Potential Molecular Dynamics: a scalable model with the accuracy of quantum mechanics
• SchNet - a deep learning architecture for molecules and materials
• Molecular Graph Convolutions: Moving Beyond Fingerprints
• Quantum-Chemical Insights from Deep Tensor Neural Networks
• Deep learning for universal linear embeddings of nonlinear dynamics
• Convolutional Networks on Graphs for Learning Molecular Fingerprints
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• VAMPnets: Deep learning of molecular kinetics
• Extended dynamic mode decomposition with dictionary learning: a data-driven adaptive spectral decomposition of the Koopman operator
• Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics
• Correlations from ion-pairing and the Nernst-Einstein equation
• Atom2Vec: learning atoms for materials discovery
• Hierarchical Visualization of Materials Space with Graph Convolutional Neural Networks
• Deep Generative Markov State Models
• Unsupervised landmark analysis for jump detection in molecular dynamics simulations
• Inference of hidden structures in complex physical systems by multi-scale clustering

Cited by in corpus (21)

• Big-Data Science in Porous Materials: Materials Genomics and Machine Learning
• Predicting materials properties without crystal structure: Deep representation learning from stoichiometry
• Deep Potential generation scheme and simulation protocol for the Li10GeP2S12-type superionic conductors
• Machine Learning on Neutron and X-Ray Scattering
• Advances of Machine Learning in Materials Science: Ideas and Techniques
• PANNA: Properties from Artificial Neural Network Architectures
• GraphVAMPNet, using graph neural networks and variational approach to markov processes for dynamical modeling of biomolecules
• Differentiable Molecular Simulations for Control and Learning
• First-Principles Prediction of Electronic Transport in Experimental Semiconductor Heterostructures via Physics-Based Machine Learning
• Ligand-induced protein dynamics differences correlate with protein-ligand binding affinities: An unsupervised deep learning approach
• Exploring particle dynamics during self-organization processes via rotationally invariant latent representations
• TimeSOAP: Tracking high-dimensional fluctuations in complex molecular systems via time-variations of SOAP spectra
• Deep learning Markov and Koopman models with physical constraints
• Exploring glassy dynamics with Markov state models from graph dynamical neural networks
• Elucidating Interfacial Dynamics of Ti-Al Systems Using Molecular Dynamics Simulation and Markov State Modeling
• LeaPP: Learning Pathways to Polymorphs through machine learning analysis of atomic trajectories
• Manifold Coordinates with Physical Meaning
• Internal-Coordinate Density Modelling of Protein Structure: Covariance Matters
• Progress in deep Markov State Modeling: Coarse graining and experimental data restraints
• Self-Supervised Decomposition, Disentanglement and Prediction of Video Sequences while Interpreting Dynamics: A Koopman Perspective
• Markov State Models for Tracking Reaction Dynamics on Catalytic Nanoparticles