Machine-learning for stellar spectroscopy: from RAVE to Gaia-RVS & 4MOST

by Dr Guillaume Guiglion (ZAH/ LSW, MPIA)

Friday 13 Oct 2023, 10:30 → 11:30 Europe/Stockholm

FC61 (AlbaNova Main Building)

In this seminar, I will present recent developments in the field of machine-learning applied to stellar spectra in the context of large scale spectroscopic surveys, such as Gaia-ESO and RAVE. I will particularly focus on Gaia DR3, which provided the community with approximately one million RVS spectra covering the CaII triplet region. One third of the spectra have a signal-to-noise ratio from 15 to 25 per pixel. I will demonstrate that precise parametrization can be achieved for such a type of dataset by using machine-learning and the full power of the Gaia data. I will present a new approach in the form of a hybrid Convolutional Neural-Network (CNN) to derive atmospheric parameters (Teff, log(g), and [M/H]) and chemical abundances ([Fe/H] and [α/M]).
Our CNN is designed to effectively combine the Gaia DR3 RVS spectra, photometry (G, Bp, Rp), parallaxes, and XP coefficients and is able to extract additional information from non-spectral inputs to   supplement the limited spectral coverage of the RVS spectra. We manage to characterize the [α/M] − [M/H] bimodality from the inner regions to the outer part of the Milky Way, which has never been characterized using RVS spectra or similar datasets. I will also discuss the benefits of using CNNs for future  large scale spectroscopic surveys such as 4MOST.