Genetic algorithms for computational materials discovery accelerated by machine learning

Materials discovery is increasingly being impelled by machine learning methods that rely on pre-existing datasets. Where datasets are lacking, unbiased data generation can be achieved with genetic algorithms. Here a machine learning model is trained on-the-fly as a computationally inexpensive energy...

Full description

Saved in:
Bibliographic Details
Published in:npj computational materials 2019-04, Vol.5 (1), Article 46
Main Authors: Jennings, Paul C., Lysgaard, Steen, Hummelshøj, Jens Strabo, Vegge, Tejs, Bligaard, Thomas
Format: Article
Language:English
Subjects:
639/301/1034/1037
639/638/563
639/925/357/354
Algorithms
Artificial intelligence
Binary alloys
Catalysts
Characterization and Evaluation of Materials
Chemistry and Materials Science
Computational Intelligence
Computer applications
Datasets
Density functional theory
Genetic algorithms
Learning algorithms
Machine learning
MATERIALS SCIENCE
Mathematical and Computational Engineering
Mathematical and Computational Physics
Mathematical Modeling and Industrial Mathematics
Nanoalloys
Nanoparticles
Search algorithms
Theoretical
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Materials discovery is increasingly being impelled by machine learning methods that rely on pre-existing datasets. Where datasets are lacking, unbiased data generation can be achieved with genetic algorithms. Here a machine learning model is trained on-the-fly as a computationally inexpensive energy predictor before analyzing how to augment convergence in genetic algorithm-based approaches by using the model as a surrogate. This leads to a machine learning accelerated genetic algorithm combining robust qualities of the genetic algorithm with rapid machine learning. The approach is used to search for stable, compositionally variant, geometrically similar nanoparticle alloys to illustrate its capability for accelerated materials discovery, e.g., nanoalloy catalysts. The machine learning accelerated approach, in this case, yields a 50-fold reduction in the number of required energy calculations compared to a traditional “brute force” genetic algorithm. This makes searching through the space of all homotops and compositions of a binary alloy particle in a given structure feasible, using density functional theory calculations.
ISSN:2057-3960
2057-3960
DOI:10.1038/s41524-019-0181-4