Deriving the star formation histories of galaxies from spectra with simulation-based inference
Published in Astronomy & Astrophysics, 2024
We explore a simulation-based inference workflow to infer posterior distributions of metallicities and star formation histories from optical absorption spectra. Trained on MILES stellar population models, we achieve R²=0.97 for the time at which a galaxy formed 50% of its stellar mass, with posterior sampling in 10⁻⁴ s — up to five orders of magnitude faster than classical methods.
Recommended citation: Iglesias-Navarro, P., Huertas-Company, M., Martín-Navarro, I., Knapen, J. H., Pernet, E. (2024). A&A, 689, A58.