(611b) Pythia: A Toolbox for Structural Analysis with Machine Learning

The recent explosion of interest and progress in machine learning (ML) methods has driven a proliferation of their application to soft matter systems. ML promises to deliver novel, automatic characterization techniques to solve previously insurmountable problems and it has already been successfully applied in several key areas for both disordered and ordered materials. However, researchers attempting to utilize ML methods on new systems often encounter challenges when the most appropriate data representation for their problem of interest is still unknown. To help alleviate this problem, we present Pythia, an open-source python library for generating numerical descriptions of particle configurations. Pythia allows users to select from a palette of descriptors ranging in complexity from simple to sophisticated. We demonstrate how Pythia can be combined with standard ML methods to identify complex structures, create phase diagrams, analyze crystal grains, and more—all in a high-throughput manner.