Graph Dynamical Networks for Unsupervised Learning of Atomic Scale Dynamics in Materials
This provides an automated tool for materials scientists to analyze atomic dynamics in functional materials, addressing global energy and environmental challenges, though it is incremental as it applies existing graph-based methods to a new domain.
The authors tackled the problem of understanding atomic-scale dynamics in complex materials, such as lithium ions in electrolytes, by developing graph dynamical networks, an unsupervised learning method that extracts important dynamical information from molecular dynamics simulations for multi-component amorphous systems.
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.