Multifidelity Linear Regression for Scientific Machine Learning from Scarce Data

Proposing a multifidelity training approach for scientific machine learning to improve model accuracy and robustness with limited high-fidelity data.