Unsupervised topological learning for identification of atomic structures

Abstract

We propose an unsupervised learning methodology with descriptors based on Topological Data Analysis (TDA) concepts to describe the local structural properties of materials at the atomic scale. Based only on atomic positions and without a priori knowledge, our method allows for an autonomous identification of clusters of atomic structures through a Gaussian mixture model. We apply successfully this approach to the analysis of elemental Zr in the crystalline and liquid states as well as homogeneous nucleation events under deep undercooling conditions. This opens the way to deeper and autonomous study of complex phenomena in materials at the atomic scale.

References in corpus (12)

• Accurate determination of crystal structures based on averaged local bond order parameters
• Crystal Nucleation in Liquids: Open Questions and Future Challenges in Molecular Dynamics Simulations
• Hierarchical structures of amorphous solids characterized by persistent homology
• Bond orientational ordering in liquids: Towards a unified description of water-like anomalies, liquid-liquid transition, glass transition, and crystallization
• Confidence sets for persistence diagrams
• Unsupervised machine learning in atomistic simulations, between predictions and understanding
• Structure of the First and Second Neighbor Shells of Water: Quantitative Relation with Translational and Orientational Order
• A parameter-free, solid-angle based, nearest-neighbor algorithm
• Crystal nucleation as the ordering of multiple order parameters
• Unsupervised learning for local structure detection in colloidal systems
• Assessing the structural heterogeneity of supercooled liquids through community inference
• The Early Crystal Nucleation Process in Hard Spheres shows Synchronised Ordering and Densification