Detection of Topological Materials with Machine Learning

TL;DR

This paper demonstrates how machine learning, specifically gradient boosted trees, can rapidly and accurately predict the topological nature of materials, significantly speeding up the discovery process compared to traditional ab-initio calculations.

Contribution

The authors develop a machine learning model that predicts topological materials with 90% accuracy, highlighting key properties and error sources, advancing materials discovery methods.

Findings

Achieved 90% prediction accuracy for topological materials.

Identified properties that influence topological classification.

Analyzed sources of model errors and potential improvements.

Abstract

Databases compiled using ab-initio and symmetry-based calculations now contain tens of thousands of topological insulators and topological semimetals. This makes the application of modern machine learning methods to topological materials possible. Using gradient boosted trees, we show how to construct a machine learning model which can predict the topology of a given existent material with an accuracy of 90%. Such predictions are orders of magnitude faster than actual ab-initio calculations. Through extensive testing of different models we determine which properties help detect topological materials. We identify the sources of our model's errors and we discuss approaches to overcome them.