Stellar Structure and Evolution
Research Projects
AI-Powered Modeling of Stellar Atmospheres
In recent years a major effort was dedicated to collecting large data sets of high-quality spectra of stars in the Milky Way and to improving data analysis pipelines with machine learning and artificial intelligence. This allowed us to efficiently obtain atmospheric parameters and chemical abundances for hundreds of thousand of stars. However, less effort was dedicated to improving the stellar atmospheric models these analyses depend on. The aim of this project is to improve the understanding of stellar atmospheres by using AI tools to improve the modeling of stellar atmospheres with current technologies.
People
The Gaia Benchmark Stars
Stellar spectroscopy is the technique to determine stellar parameters and chemical abundances of stars, and relies on the modeling of stellar atmospheres which are based on simplistic (therefore uncertain) assumptions. The Gaia Benchmark Stars are well-known stars from which information about their structure beyond the spectra can be obtained (interferometric and astrometric measurements). These stars have become the calibration pillars for Gaia and for its complementary spectroscopic surveys. At UDP we assemble them, observe them, perform high resolution spectroscopy, and assess the quality of lines in both the optical and the infrared spectral range.