The Gaia-ESO Survey: Preparing the ground for 4MOST & WEAVE galactic surveys. Chemical evolution of lithium with machine learning

Abstract

With its origin coming from several sources (Big Bang, stars, cosmic rays) and given its strong depletion during its stellar lifetime, the lithium element is of great interest as its chemical evolution in the Milky Way is not well understood at present. To help constrain stellar and galactic chemical evolution models, numerous and precise lithium abundances are necessary for a large range of evolutionary stages, metallicities, and Galactic volume. In the age of stellar parametrization on industrial scales, spectroscopic surveys such as APOGEE, GALAH, RAVE, and LAMOST have used data-driven methods to rapidly and precisely infer stellar labels (atmospheric parameters and abundances). To prepare the ground for future spectroscopic surveys such as 4MOST and WEAVE, we aim to apply machine learning techniques to lithium measurements and analyses. We trained a convolution neural network (CNN), coupling Gaia-ESO Survey iDR6 stellar labels (Teff, log(g), [Fe/H], and A(Li)) and GIRAFFE HR15N spectra, to infer the atm parameters and lithium abundances for ~40,000 stars. We show that the CNN properly learns the physics of the stellar labels, from relevant spectral features through a broad range of evolutionary stages and stellar parameters. The Li feature at 6707.8 A is successfully singled out by our CNN, among the thousands of lines. Rare objects such as Li-rich giants are found in our sample. This level of performance is achieved thanks to a meticulously built, high-quality, and homogeneous training sample. The CNN approach is very well adapted for the next generations of spectroscopic surveys aimed at studying (among other elements) lithium, such as the 4MIDABLE-LR/HR (4MOST Milky Way disk and bulge low- and high-resolution) surveys. In this context, the caveats of ML applications should be appropriately investigated, along with the realistic label uncertainties and upper limits for abundances.

arXiv commentA&A, accepted 05 December 2022, 23 pages, 23 figures. Codes and trained models available at https://github.com/SamirNepal/Li\_CNN\_2022

References in corpus (79)

• Array Programming with NumPy
• The Gaia mission
• Gaia Early Data Release 3: The astrometric solution
• The GALAH+ Survey: Third Data Release
• The Primordial Lithium Problem
• Precision big bang nucleosynthesis with improved Helium-4 predictions
• The Cannon: A data-driven approach to stellar label determination
• NLTE line formation of Fe for late-type stars. I. Standard stars with 1D and <3D> model atmospheres
• Departures from LTE for neutral Li in late-type stars
• Gaia FGK Benchmark Stars - Metallicity
• Gaia FGK Benchmark Stars: Effective temperatures and surface gravities
• Thermohaline instability and rotation-induced mixing. III - Grid of stellar models and asymptotic asteroseismic quantities from the pre-main sequence up to the AGB for low- and intermediate-mass stars at various metallicities
• The primordial lithium abundance
• The Payne: self-consistent ab initio fitting of stellar spectra
• Gaia-ESO Survey: The analysis of high-resolution UVES spectra of FGK-type stars
• Lithium abundances in nearby FGK dwarf and subgiant stars: internal destruction, Galactic chemical evolution, and exoplanets
• Abundance Estimates for 16 Elements in 6 Million Stars from LAMOST DR5 Low-Resolution Spectra
• Hunting for open clusters in \textit{Gaia} DR2: $582$ new OCs in the Galactic disc
• Estimating stellar birth radii and the time evolution of the Milky Way's ISM metallicity gradient
• Age dissection of the Milky Way discs: red giants in the Kepler field
• The Gaia-ESO Public Spectroscopic Survey: Implementation, data products, open cluster survey, science, and legacy
• Finding Strong Gravitational Lenses in the Kilo Degree Survey with Convolutional Neural Networks
• Atomic data for the Gaia-ESO Survey
• Gaia FGK benchmark stars: abundances of alpha and iron-peak elements
• Li abundances in F stars: planets, rotation and galactic evolution
• The Gaia FGK Benchmark Stars - High resolution spectral library
• Lithium Abundances of the Local Thin Disk Stars
• The Gaia-ESO Public Spectroscopic Survey: Motivation, implementation, GIRAFFE data processing, analysis, and final data products
• An Application of Deep Neural Networks in the Analysis of Stellar Spectra
• Automated Classification of Stellar Spectra. II: Two-Dimensional Classification with Neural Networks and Principal Components Analysis
• Non-local thermodynamic equilibrium stellar spectroscopy with 1D and <3D> models - I. Methods and application to magnesium abundances in standard stars
• On Lithium-Rich Red Giants. I. Engulfment of Sub-Stellar Companions
• Solar oxygen abundance
• 3D NLTE spectral line formation of lithium in late-type stars
• Dissecting stellar chemical abundance space with t-SNE
• Early optical spectra of nova V1369 Cen show presence of Lithium
• The stellar origin of 7Li - Do AGB stars contribute a substantial fraction of the local Galactic lithium abundance?
• Physical Parameterization of Stellar Spectra: The Neural Network Approach
• The AMBRE Project: Constraining the lithium evolution in the Milky Way
• The Gaia-ESO Survey: calibration strategy
• Lithium-rich Giants in LAMOST Survey. I. The Catalog
• Detecting outliers in astronomical images with deep generative networks
• Exploring the production and depletion of lithium in the Milky Way stellar disk
• The Gaia-ESO Survey: Galactic evolution of lithium from iDR6
• The Gaia-ESO survey: Mixing processes in low-mass stars traced by lithium abundance in cluster and field stars
• Prospects for Measuring Abundances of >20 Elements with Low-resolution Stellar Spectra
• Gaia-ESO Survey: The analysis of pre-main sequence stellar spectra
• Very metal-poor stars observed by the RAVE survey
• The Gaia-ESO Survey: Empirical determination of the precision of stellar radial velocities and projected rotation velocities
• ${}^7$Li evolution in the thin and thick disks of the Milky Way
• The Gaia -ESO Survey: Lithium enrichment histories of the Galactic thick and thin disc
• Detection of Gravitational Waves Using Bayesian Neural Networks
• Light Elements in the Universe
• The Gaia-ESO Survey: Target selection of open cluster stars
• Li Evolution and the Open Cluster NGC 6819: A Correlation Between Li Depletion and Spindown in Dwarfs More Massive Than the F-Dwarf Li-Dip
• The Gaia-ESO Survey: properties of newly discovered Li-rich giants
• The GALAH Survey: A new constraint on cosmological lithium and Galactic lithium evolution from warm dwarf stars
• Empirical classification of VLT/Giraffe stellar spectra in the wavelength range 6440-6810 A in the gamma Vel cluster, and calibration of spectral indices
• The Gaia-ESO Survey: the selection function of the Milky Way field stars
• The AMBRE project: a study of Li evolution in the Galactic thin and thick discs
• The Gaia-ESO Survey: Galactic evolution of lithium at high metallicity
• Explaining the decrease in ISM lithium at super-solar metallicities in the solar vicinity
• Cycle-StarNet: Bridging the gap between theory and data by leveraging large datasets
• Measuring oxygen abundances from stellar spectra without oxygen lines
• The RAdial Velocity Experiment (RAVE): Parameterisation of RAVE spectra based on convolutional neural networks
• Supernova Neutrino Process of Li and B Revisited
• Tracking the evolution of lithium in giants using asteroseismology: Super-Li-rich stars are almost exclusively young red-clump stars
• A NLTE line formation for neutral and singly-ionised calcium in model atmospheres of B-F stars
• The Gaia-ESO survey: Lithium abundances in open cluster Red Clump stars
• Assessing the performance of LTE and NLTE synthetic stellar spectra in a machine learning framework
• The Gaia-ESO Survey: Lithium measurements and new curves of growth
• The Gaia-ESO Survey: Chemical evolution of Mg and Al in the Milky Way with Machine-Learning
• The Gaia-ESO Survey: an extremely Li-rich giant in the globular cluster NGC 1261
• Li-rich Giants in LAMOST Survey. III. The statistical analysis of Li-rich giants
• The behaviour of lithium at high metallicity in the Milky Way -- Selection effects in the samples and the possible role of atomic diffusion
• The CoRoT-GES Collaboration. Improving Red Giants spectroscopic surface gravity and abundances with asteroseismology
• Lithium in red giants: the roles of the He-core flash and the luminosity bump
• The GALAH survey: Characterization of emission-line stars with spectral modelling using autoencoders
• The Gaia-ESO Survey: Probing the lithium abundances in old metal-rich dwarf stars in the Solar vicinity

Cited by in corpus (7)

• The Gaia-ESO Survey: Chemical evolution of Mg and Al in the Milky Way with Machine-Learning
• Galaxy Spectra neural Network (GaSNet). II. Using Deep Learning for Spectral Classification and Redshift Predictions
• Evolution of lithium in the disc of the Galaxy and the role of novae
• Advanced Classification of Hot Subdwarf Binaries Using Artificial Intelligence Techniques and Gaia DR3 data
• The GALAH survey: Improving chemical abundances using star clusters
• Playing CHESS with stars. I. Search for similar stars in large spectroscopic data sets
• Exploring Late Stellar Evolution in the Era of Large Surveys: Machine Learning Prospects for Hot Subdwarfs and White Dwarfs