Machine Learning Based System for Automated Estimation of Crystal Unit Cell Parameters from X-ray Diffraction Powder Patterns (MLCell) v0.1

Abstract

This software uses machine learning models to perform inference on X-ray diffraction powder patterns, output from a diffractometer, to estimate all 6 unit cell parameters of crystalline lattices. Novel aspects of this software originated from the incorporation of concepts from crystallography to an ML-based approach; The resulting method achieves new performance capabilities without significant loss of accuracy and with less computationally intensive models.