(156f) Combining Molecular Dynamics Simulations and Deep Learning to Predict Interfacial Properties

Molecular dynamics (MD) simulations are a powerful tool for analyzing the properties and behavior of materials in solvent environments by explicitly representing atomistic degrees of freedom and accounting for intermolecular interactions. However, MD simulations can require lengthy runtimes that limit the high-throughput screening of system properties. Moreover, the analysis of MD simulation output typically requires the implementation of analytical techniques that rely on human insight to relate the properties of a statistical ensemble to relevant macroscopic properties. Alternatively, machine learning has emerged as a valuable tool to interpret molecular simulation data. Here, we present computational techniques that use machine learning to analyze molecular simulations of complex solvent environments. In particular, we focus on quantifying the hydrophobicity of functionalized interfaces by utilizing deep learning to analyze MD-derived interfacial water configurations. By selecting an appropriate representation of the MD output data, we demonstrate that a convolutional neural network (CNN) can infer spatial correlations between solvent molecules and map these features to thermodynamic observables. Once trained, we find that the CNN can predict dewetting free energies with minimal simulation input, facilitating the rapid screening of interfacial hydrophobicity for a range of surface compositions. We further use feature visualization methods to provide a physical interpretation of features identified by the neural network and compare these features to typical observables used to describe interfacial hydrophobicity (i.e., water structure). This work demonstrates how deep learning methods can complement conventional MD simulations to facilitate the screening of interfacial properties.