Machine-learning methods for stellar model calibration with large datasets

In mid-2022, the European Space Agency (ESA) will release new results from the Gaia mission. This third data release (DR3) will be the first one to include mass, radius and luminosity estimates for eclipsing binaries stars based on Gaia data. Eclipsing binary stars are our only source of model-independent masses and radii for normal stars, so they are essential for testing and calibrating stellar models, e.g. the models that will be used by ESA's PLATO mission to characterise planet host stars. The final Gaia data release (DR4) will contain mass, radius and luminosity measurements for thousands of eclipsing binaries. The aim of this PhD project will be to develop new techniques to compare these fundamental measurements from Gaia, complemented by observations from other surveys and ground-based follow-up observations, to a wide range of stellar models. This is a well-known method for calibrating unknown parameters in stellar models but has never been attempted on such a large scale before. This will require a new approach based on machine learning algorithms and advanced statistical techniques. 

Please quote reference FNS 2021-19 when applying.