An explainable ML model for binary LJ fluids

Lennard-Jones (LJ) fluids serve as an important theoretical framework for understanding molecular interactions. Binary LJ fluids, where two distinct species of particles interact based on the LJ potential, exhibit rich phase behavior and provide valuable insights into complex fluid mixtures. Here we report the construction and utility of a machine learning (ML) model for binary LJ fluids, focusing on their effectiveness in predicting radial distribution functions (RDFs) across a range of conditions. The RDFs of a binary mixture with varying compositions and temperatures are collected from molecular dynamics (MD) simulations to establish and validate the ML model. In this ML pipeline, RDFs are discretized in order to reduce the output dimensions of the model. This, in turn, improves the efficacy, and reduces the complexity of an ML RDF model. The model is shown to predict RDFs for many unknown mixtures very accurately. We further report the model's ability to extrapolate within the multiphase compositional - temperature space of the binary LJ fluid. Our ML model suggests that the particle size ratio exerts a dominant influence on the microstructure of a binary mixture. This work highlights the limitations of the ML model when it encounters new regimes governed by different underlying physics that are not represented in the training data.