This paper presents a multi-fidelity machine learning approach to high-throughput materials screening, using a multi-output Gaussian process to fuse experimental and computational data. This approach avoids challenges associated with traditional computational funnels by dynamically learning relationships between methods. Evaluation on three materials design problems shows a three-fold reduction in optimization cost compared to other methods.
Publisher
npj Computational Materials
Published On
Dec 19, 2022
Authors
Clyde Fare, Peter Fenner, Matthew Benatan, Alessandro Varsi, Edward O. Pyzer-Knapp
Tags
multi-fidelity
machine learning
materials screening
experimental data
computational data
Gaussian process
optimization
Related Publications
Explore these studies to deepen your understanding of the subject.
