J/A+A/671/A61 Lithium with Machine-Learning (Nepal+, 2023)
The Gaia-ESO Survey: Preparing the ground for 4MOST & WEAVE galactic surveys.
Chemical evolution of lithium with machine learning.
Nepal S., Guiglion G., de Jong R.S., Valentini M., Chiappini C.,
Steinmetz M., Ambrosch M., Pancino E., Jeffries R.D., Bensby T., Romano D.,
Smiljanic R., Dantas M.L.L., Gilmore G., Randich S., Bayo A., Bergemann M.,
Franciosini E., Jimenez Esteban F., Jofre P., Morbidelli L., Sacco G.G.,
Tautvaisiene G., Zaggia S.
<Astron. Astrophys. 671, A61 (2023)>
=2023A&A...671A..61N 2023A&A...671A..61N (SIMBAD/NED BibCode)
ADC_Keywords: Surveys ; Stars, standard ; Abundances ; Spectroscopy ;
Effective temperatures ; Optical
Keywords: techniques: spectroscopic - methods: data analysis - surveys -
stars: fundamental parameters - stars: abundances -
Galaxy: stellar content
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 atmospheric parameters and lithium
abundances for 40000 stars. The CNN architecture and accompanying
notebooks are available online via GitHub at
https://github.com/SamirNepal.
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 lithium feature at
6707.8Å, is successfully singled out by our CNN, among the
thousands of lines in the GIRAFFE HR15N setup. Rare objects such as
lithium-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 at
low and high resolution) surveys. In this context, the caveats of
machine-learning applications should be appropriately investigated,
along with the realistic label uncertainties and upper limits for
abundances.
Description:
This catalog includes atmospheric parameters and lithium abundances,
along with their uncertainties, for 40150 GIRAFFE HR15N spectra,
parametrized using Convolutional Neural Networks. The catalog also
contains, the spectra name and GES identifier, the signal-to-noise
ratio of the spectra and flags indicating which labels are within the
training set limits. The training data for this machine learning
application were obtained from the Gaia-ESO survey. The accompanying
paper describes the training data, the CNN method, and presents
extensive validation using benchmarks and external catalogs, as well
as two science applications of the catalog.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table1.dat 263 40150 Catalog of CNN result
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 16 A16 --- CName GES ID (HHMMSSss+DDMMSSs)
18- 52 A35 --- spName Name of the spectrum
54- 71 F18.13 K Teff Effective temperature
73- 89 F17.13 K e_Teff Uncertainty of Teff
91 A1 --- f_Teff [01] Boundary flag for Teff (1)
93-111 F19.17 [cm/s2] logg Surface gravity
113-136 E24.16 [cm/s2] e_logg Uncertainty of log(g)
138 A1 --- f_logg [01] Boundary flag for log(g) (1)
140-163 E24.16 --- [Fe/H] Metallicity [Fe/H]
165-188 E24.16 --- e_[Fe/H] Uncertainty of [Fe/H]
190 I1 --- f_[Fe/H] [0/1] Boundary flag for [Fe/H] (1)
192-214 F23.20 --- A(Li) Lithium abundance
216-239 E24.16 --- e_A(Li) Uncertainty of A(Li)
241 I1 --- f_A(Li) [0/1] Boundary flag for A(Li) (1)
243-263 F21.16 pix-1 snr Signal-to-noise ratio
--------------------------------------------------------------------------------
Note (1): Flag as follows:
1 = inside training limit
0 = outside the training limit
--------------------------------------------------------------------------------
Acknowledgements:
Samir Nepal, snepal(at)aip.de
(End) Patricia Vannier [CDS] 28-Dec-2022