Multifidelity Linear Regression

Multifidelity linear regression leverages data of varying accuracy (fidelities) to build more robust and accurate predictive models, particularly when high-fidelity data is scarce and expensive to obtain. Current research focuses on applying this technique with various model architectures, including linear regression, deep Gaussian processes, and convolutional neural networks, often incorporating gradient information to improve performance. This approach significantly enhances the efficiency and reliability of scientific machine learning by reducing model variance and improving predictions, impacting fields ranging from quantum chemistry to fluid dynamics simulations.