Transferring spectroscopic stellar labels to 217 million Gaia DR3 XP stars with SHBoost

With Gaia Data Release 3 (DR3), new and improved astrometric, photometric, and spectroscopic measurements for 1.8 billion stars
are available. Alongside this wealth of new data, however, come challenges in finding increasingly efficient and accurate computational
methods to use for analysis. In this paper we explore the feasibility of using machine-learning regression as a method of extracting
basic stellar parameters and line-of-sight extinctions, given spectro-photometric data. To this end, we build a stable gradient-boosted
random-forest regressor (xgboost), trained on spectroscopic data, capable of producing output parameters with reliable uncertainties
from Gaia DR3 data (most notably the low-resolution XP spectra) without ground-based spectroscopic observations. Using Shapley
additive explanations, we are also able to interpret how the predictions for each star are influenced by each data feature. For the
training and testing of the network, we use high-quality parameters obtained from the StarHorse code for a sample of around eight
million stars observed by major spectroscopic surveys (APOGEE, GALAH, LAMOST, RAVE, SEGUE, and GES), complemented
by curated samples of hot stars, very metal-poor stars, white dwarfs, and hot sub-dwarfs. The training data cover the whole sky, all
Galactic components, and almost the full magnitude range of the Gaia DR3 XP sample of more than 217 million objects that also
have parallaxes. We achieve median uncertainties (at G ≈ 16) of 0.20 mag in V-band extinction, 0.01 dex in logarithmic effective temperature, 0.20 dex in surface gravity, 0.18 dex in metallicity, and 12% in mass (over the full Gaia DR3 XP sample, with considerable
variations in precision as a function of magnitude and stellar type). We succeed in predicting competitive results based on Gaia DR3
XP spectra compared to classical isochrone or spectral-energy distribution fitting methods we employed in earlier work, especially
for the parameters AV , Teff, and metallicity. Finally, we showcase some potential applications of this new catalogue (e.g. extinction
maps, metallicity trends in the Milky Way, extended maps of young massive stars, metal-poor stars, and metal-rich stars).