J/A+A/673/A105 Deep Transfer Learning of Teff and [M/H] (Bello-Garcia+, 2023)
The CARMENES search for exoplanets around M dwarfs:
A deep transfer learning method to determine Teff and [M/H] of target stars.
Bello-Garcia A., Passegger V.M., Ordieres-Mere J., Schweitzer A.,
Caballero J.A., Gonzalez-Marcos A., Ribas I., Reiners A., Quirrenbach A.,
Amado P.J., Bejar V.J.S., Cifuentes C., Henning T., Kaminski A., Luque R.,
Montes D., Morales J.C., Pedraz S., Tabernero H.M., Zechmeister M.
<Astron. Astrophys. 673, A105 (2023)>
=2023A&A...673A.105B 2023A&A...673A.105B (SIMBAD/NED BibCode)
ADC_Keywords: Stars, M-type ; Abundances, [Fe/H] ; Effective temperatures ;
Optical
Keywords: methods: data analysis - techniques: spectroscopic -
stars: fundamental parameters - stars: late-type - stars: low-mass
Abstract:
The large amounts of astrophysical data being provided by existing and
future instrumentation require efficient and fast analysis tools.
Transfer learning is a new technique promising higher accuracy in the
derived data products, with information from one domain being
transferred to improve the accuracy of a neural network model in
another domain. In this work, we demonstrate the feasibility of
applying the deep transfer learning (DTL) approach to high-resolution
spectra in the framework of photospheric stellar parameter
determination. To this end, we used 14 stars of the CARMENES survey
sample with interferometric angular diameters to calculate the
effective temperature, as well as six M dwarfs that are common proper
motion companions to FGK-type primaries with known metallicity. After
training a deep learning (DL) neural network model on synthetic
PHOENIX-ACES spectra, we used the internal feature representations
together with those 14+6 stars with independent parameter measurements
as a new input for the transfer process. We compare the derived
stellar parameters of a small sample of M dwarfs kept out of the
training phase with results from other methods in the literature.
Assuming that temperatures from bolometric luminosities and
interferometric radii and metallicities from FGK+M binaries are
sufficiently accurate, DTL provides a higher accuracy than our
previous state-of-the-art DL method (mean absolute differences improve
by 20K for temperature and 0.2dex for metallicity from DL to DTL
when compared with reference values from interferometry and FGK+M
binaries). Furthermore, the machine learning (internal) precision of
DTL also improves as uncertainties are five times smaller on average.
These results indicate that DTL is a robust tool for obtaining M-dwarf
stellar parameters comparable to those obtained from independent
estimations for well-known stars.
Description:
Teff and [M/H] values for CARMENES stars estimated by the DTL method.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
tablea1.dat 50 286 Teff and [M/H] values for CARMENES stars
estimated by the DTL method
--------------------------------------------------------------------------------
See also:
J/A+A/577/A128 : CARMENES input catalog of M dwarfs I (Alonso-Floriano+, 2015)
J/A+A/612/A49 : 324 CARMENES M dwarfs velocities (Reiners+, 2018)
J/A+A/615/A6 : Photospheric parameters of CARMENES stars (Passegger+, 2018)
J/A+A/621/A126 : CARMENES input catalogue of M dwarfs. IV. (Diez Alonso+ 2019)
J/A+A/625/A68 : Radii and masses of the CARMENES targets (Schweitzer+, 2019)
J/A+A/642/A22 : DL fundamental parameters of target stars (Passegger+, 2020)
J/A+A/658/A194 : Metallicities in M dwarfs (Passegger+, 2022)
Byte-by-byte Description of file: tablea1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 11 A11 --- Karmn CARMENES identifier (JHHMMm+DDdA)
13- 17 A5 --- Notes Additional note (1)
19- 22 I4 K Teff Effective temperature
24- 27 I4 K e_Teff Delta Teff for -1 sigma
29- 32 I4 K E_Teff Delta Teff for +1 sigma
34- 38 F5.2 [-] [M/H] Metallicity
41- 44 F4.2 [-] e_[M/H] ?=- Delta Metallicity for -1 sigma
47- 50 F4.2 [-] E_[M/H] ?=- Delta Metallicity for +1 sigma
--------------------------------------------------------------------------------
Note (1): Note as follows:
1 = Stars used for training the DTL Teff model
2 = Stars used for training the DTL and [M/H] model
3 = to stars with rotational velocity 3sini>10km/s.
Some 1sigma uncertainties and quantiles for stars marked 1 and 2 are missing
because the different predictions were too close. In these few cases,
the quantiles could be substituted by the standard deviation.
--------------------------------------------------------------------------------
Acknowledgements:
Antonio Bello-Garcia, abello(at)uniovi.es
(End) Patricia Vannier [CDS] 14-Apr-2023