Unsupervised topological learning for identification of atomic structures

Unsupervised topological learning for identification of atomic structures

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 identifi...

Full description

Bibliographic Details
Main Authors: Becker, S. (Author), Devijver, E. (Author), Jakse, N. (Author), Molinier, R. (Author)
Format: Article
Language:English
Published: American Physical Society 2022
Subjects:
Atomic positions
Atomic scale
Atoms
Crystalline state
Descriptors
Gaussian distribution
Gaussian Mixture Model
Homogeneous nucleation
Liquid state
Local structural properties
Nucleation
Priori knowledge
Topological data analysis
Topology
Online Access:View Fulltext in Publisher
Description
Summary: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. © 2022 American Physical Society.
ISBN:24700045 (ISSN)
DOI:10.1103/PhysRevE.105.045304

Similar Items