The Gaia-ESO Survey: Chemical evolution of Mg and Al in the Milky Way with machine learning
Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Standard
The Gaia-ESO Survey : Chemical evolution of Mg and Al in the Milky Way with machine learning. / Ambrosch, M.; Guiglion, G.; Mikolaitis, S.; Chiappini, C.; Tautvaisiene, G.; Nepal, S.; Gilmore, G.; Randich, S.; Bensby, T.; Bayo, A.; Bergemann, M.; Morbidelli, L.; Pancino, E.; Sacco, G. G.; Smiljanic, R.; Zaggia, S.; Jofre, P.; Jimenez-Esteban, F. M.
I: Astronomy & Astrophysics, Bind 672, A46, 27.03.2023.Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - The Gaia-ESO Survey
T2 - Chemical evolution of Mg and Al in the Milky Way with machine learning
AU - Ambrosch, M.
AU - Guiglion, G.
AU - Mikolaitis, S.
AU - Chiappini, C.
AU - Tautvaisiene, G.
AU - Nepal, S.
AU - Gilmore, G.
AU - Randich, S.
AU - Bensby, T.
AU - Bayo, A.
AU - Bergemann, M.
AU - Morbidelli, L.
AU - Pancino, E.
AU - Sacco, G. G.
AU - Smiljanic, R.
AU - Zaggia, S.
AU - Jofre, P.
AU - Jimenez-Esteban, F. M.
PY - 2023/3/27
Y1 - 2023/3/27
N2 - Context. To take full advantage of upcoming large-scale spectroscopic surveys, it will be necessary to parameterize millions of stellar spectra in an efficient way. Machine learning methods, especially convolutional neural networks (CNNs), will be among the main tools geared at achieving this task.Aims. We aim to prepare the groundwork for machine learning techniques for the next generation of spectroscopic surveys, such as 4MOST and WEAVE. Our goal is to show that CNNs can predict accurate stellar labels from relevant spectral features in a physically meaningful way. The predicted labels can be used to investigate properties of the Milky Way galaxy.Methods. We built a neural network and trained it on GIRAFFE spectra with their associated stellar labels from the sixth internal Gaia-ESO data release. Our network architecture contains several convolutional layers that allow the network to identify absorption features in the input spectra. The internal uncertainty was estimated from multiple network models. We used the t-distributed stochastic neighbor embedding tool to remove bad spectra from our training sample.Results. Our neural network is able to predict the atmospheric parameters T-eff and log(g) as well as the chemical abundances [Mg/Fe], [Al/Fe], and [Fe/H] for 36 904 stellar spectra. The training precision is 37 K for T-eff, 0.06 dex for log(g), 0.05 dex for [Mg/Fe], 0.08 dex for [Al/Fe], and 0.04 dex for [Fe/H]. Network gradients reveal that the network is inferring the labels in a physically meaningful way from spectral features. We validated our methodology using benchmark stars and recovered the properties of different stellar populations in the Milky Way galaxy.Conclusions. Such a study provides very good insights into the application of machine learning for the analysis of large-scale spectroscopic surveys, such as WEAVE and 4MOST Milky Way disk and bulge low- and high-resolution (4MIDABLE-LR and -HR). The community will have to put substantial efforts into building proactive training sets for machine learning methods to minimize any possible systematics.
AB - Context. To take full advantage of upcoming large-scale spectroscopic surveys, it will be necessary to parameterize millions of stellar spectra in an efficient way. Machine learning methods, especially convolutional neural networks (CNNs), will be among the main tools geared at achieving this task.Aims. We aim to prepare the groundwork for machine learning techniques for the next generation of spectroscopic surveys, such as 4MOST and WEAVE. Our goal is to show that CNNs can predict accurate stellar labels from relevant spectral features in a physically meaningful way. The predicted labels can be used to investigate properties of the Milky Way galaxy.Methods. We built a neural network and trained it on GIRAFFE spectra with their associated stellar labels from the sixth internal Gaia-ESO data release. Our network architecture contains several convolutional layers that allow the network to identify absorption features in the input spectra. The internal uncertainty was estimated from multiple network models. We used the t-distributed stochastic neighbor embedding tool to remove bad spectra from our training sample.Results. Our neural network is able to predict the atmospheric parameters T-eff and log(g) as well as the chemical abundances [Mg/Fe], [Al/Fe], and [Fe/H] for 36 904 stellar spectra. The training precision is 37 K for T-eff, 0.06 dex for log(g), 0.05 dex for [Mg/Fe], 0.08 dex for [Al/Fe], and 0.04 dex for [Fe/H]. Network gradients reveal that the network is inferring the labels in a physically meaningful way from spectral features. We validated our methodology using benchmark stars and recovered the properties of different stellar populations in the Milky Way galaxy.Conclusions. Such a study provides very good insights into the application of machine learning for the analysis of large-scale spectroscopic surveys, such as WEAVE and 4MOST Milky Way disk and bulge low- and high-resolution (4MIDABLE-LR and -HR). The community will have to put substantial efforts into building proactive training sets for machine learning methods to minimize any possible systematics.
KW - Galaxy:abundances
KW - Galaxy:stellar content
KW - stars:abundances
KW - techniques:spectroscopic
KW - methods:data analysis
KW - GALACTIC DISK
KW - STARS
KW - ABUNDANCES
KW - CLASSIFICATION
KW - SPECTROSCOPY
KW - LAMOST
KW - GALAXY
KW - THICK
U2 - 10.1051/0004-6361/202244766
DO - 10.1051/0004-6361/202244766
M3 - Journal article
VL - 672
JO - Astronomy & Astrophysics
JF - Astronomy & Astrophysics
SN - 0004-6361
M1 - A46
ER -
ID: 346955382