An Application of Deep Neural Networks in the Analysis of Stellar Spectra

Abstract

Spectroscopic surveys require fast and efficient analysis methods to maximize their scientific impact. Here we apply a deep neural network architecture to analyze both SDSS-III APOGEE DR13 and synthetic stellar spectra. When our convolutional neural network model (StarNet) is trained on APOGEE spectra, we show that the stellar parameters (temperature, gravity, and metallicity) are determined with similar precision and accuracy as the APOGEE pipeline. StarNet can also predict stellar parameters when trained on synthetic data, with excellent precision and accuracy for both APOGEE data and synthetic data, over a wide range of signal-to-noise ratios. In addition, the statistical uncertainties in the stellar parameter determinations are comparable to the differences between the APOGEE pipeline results and those determined independently from optical spectra. We compare StarNet to other data-driven methods; for example, StarNet and the Cannon 2 show similar behaviour when trained with the same datasets, however StarNet performs poorly on small training sets like those used by the original Cannon. The influence of the spectral features on the stellar parameters is examined via partial derivatives of the StarNet model results with respect to the input spectra. While StarNet was developed using the APOGEE observed spectra and corresponding ASSET synthetic data, we suggest that this technique is applicable to other wavelength ranges and other spectral surveys.

arXiv commentAccepted 19 December 2017 to MNRAS. 16 pages, 14 figures, 2 appendices

References in corpus (43)

• The Apache Point Observatory Galactic Evolution Experiment (APOGEE)
• The Dartmouth Stellar Evolution Database
• SEGUE: A Spectroscopic Survey of 240,000 stars with g=14-20
• Na-O Anticorrelation and HB. VII. The chemical composition of first and second-generation stars in 15 globular clusters from GIRAFFE spectra
• ASPCAP: The Apogee Stellar Parameter and Chemical Abundances Pipeline
• Abundances for Globular Cluster Giants: I. Homogeneous Metallicities for 24 Clusters
• The SEGUE Stellar Parameter Pipeline. I. Description and Initial Validation Tests
• The Thirteenth Data Release of the Sloan Digital Sky Survey: First Spectroscopic Data from the SDSS-IV Survey MApping Nearby Galaxies at Apache Point Observatory
• A Search for Stars of Very Low Metal Abundance. VI. Detailed Abundances of 313 Metal-Poor Stars
• The Data Reduction Pipeline for the Apache Point Observatory Galactic Evolution Experiment
• Abundances, Stellar Parameters, and Spectra From the SDSS-III/APOGEE Survey
• The Cannon: A data-driven approach to stellar label determination
• The Abundances of Neutron Capture Species in the Very Metal-Poor Globular Cluster M15: An Uniform Analysis of RGB and RHB Stars
• New ATLAS9 And MARCS Model Atmosphere Grids for the Apache Point Observatory Galactic Evolution Experiment (APOGEE)
• The SEGUE Stellar Parameter Pipeline. III. Comparison with High-Resolution Spectroscopy of SDSS/SEGUE Field Stars
• Gaia-ESO Survey: The analysis of high-resolution UVES spectra of FGK-type stars
• The PASTEL catalogue: 2016 version
• Chemical abundance analysis of the Open Clusters Cr110, NGC2099 (M37), NGC2420, NGC7789 and M67 (NGC2682)
• The Chemical Composition Contrast between M3 and M13 Revisited: New Abundances for 28 Giant Stars in M3
• Metal Abundances of Red Clump Stars in Open Clusters: I. NGC 6819
• High-Resolution Spectroscopy of Extremely Metal-Poor Stars from SDSS/SEGUE: I. Atmospheric Parameters and Chemical Compositions
• Exploring Anticorrelations and Light Element Variations in Northern Globular Clusters Observed by the APOGEE Survey
• Calibrations of Atmospheric Parameters Obtained from the First Year of SDSS-III APOGEE Observations
• The SEGUE Stellar Parameter Pipeline. V. Estimation of Alpha-Element Abundance Ratios From Low-Resolution SDSS/SEGUE Stellar Spectra
• Element abundances of unevolved stars in the open cluster M 67
• The Most Metal-Poor Stars. III. The Metallicity Distribution Function and CEMP Fraction
• Center-to-Limb Variation of Solar 3-D Hydrodynamical Simulations
• The chemical abundance of the very metal rich old Open Clusters NGC 6253 and NGC 6791
• The metallicity of the old open cluster NGC 6791
• Sodium and Oxygen Abundances in the Open Cluster NGC 6791 from APOGEE H-Band Spectroscopy
• The shortest period sdB plus white dwarf binary CD-30 11223 (GALEX J1411-3053)
• Physical Parameterization of Stellar Spectra: The Neural Network Approach
• The Gaia-ESO Survey: the most metal-poor stars in the Galactic bulge
• Aluminum Abundances, Deep Mixing and the Blue-Tail Second-Parameter Effect in the Globular Clusters M 3 and M 13
• The Gaia-ESO Survey: calibration strategy
• Heavy Element Dispersion in the Metal-Poor Globular Cluster M92
• An Extremely Lithium-Rich Bright Red Giant in the Globular Cluster M3
• The Cannon 2: A data-driven model of stellar spectra for detailed chemical abundance analyses
• The Gaia-ESO Survey: low-alpha element stars in the Galactic Bulge
• Chemical Abundances for Evolved Stars in M5: Lithium through Thorium
• Tidal inertial waves in the differentially rotating convective envelopes of low-mass stars - I. Free oscillation modes
• Chemical abundances in the globular clusters NGC 5024 and NGC 5466 from optical and infrared spectroscopy
• The Peculiar Globular Cluster Palomar 1 and Persistence in the SDSS-APOGEE Database

Cited by in corpus (64)

• The Challenge of Machine Learning in Space Weather Nowcasting and Forecasting
• The Payne: self-consistent ab initio fitting of stellar spectra
• Deep learning of multi-element abundances from high-resolution spectroscopic data
• Simultaneous calibration of spectro-photometric distances and the Gaia DR2 parallax zero-point offset with deep learning
• Radio Galaxy Zoo: ClaRAN - A Deep Learning Classifier for Radio Morphologies
• 3D NLTE spectral line formation of lithium in late-type stars
• Application of Convolutional Neural Networks for Stellar Spectral Classification
• SPCANet: Stellar Parameters and Chemical Abundances Network for LAMOST-II Medium Resolution Survey
• Generalizing semi-supervised generative adversarial networks to regression using feature contrasting
• J-PLUS: Identification of low-metallicity stars with artificial neural networks using SPHINX
• A Machine Learning Based Morphological Classification of 14,245 Radio AGNs Selected From The Best-Heckman Sample
• How Many Elements Matter?
• BACCHUS Analysis of Weak Lines in APOGEE Spectra (BAWLAS)
• Learning the Relationship between Galaxies Spectra and their Star Formation Histories using Convolutional Neural Networks and Cosmological Simulations
• APOGEE Net: Improving the derived spectral parameters for young stars through deep learning
• Using convolutional neural networks to predict galaxy metallicity from three-color images
• The CARMENES search for exoplanets around M dwarfs -- A deep learning approach to determine fundamental parameters of target stars
• A variational encoder-decoder approach to precise spectroscopic age estimation for large Galactic surveys
• Extracting the cold neutral medium from HI emission with deep learning: Implications for Galactic foregrounds at high latitude
• Cycle-StarNet: Bridging the gap between theory and data by leveraging large datasets
• The RAdial Velocity Experiment (RAVE): Parameterisation of RAVE spectra based on convolutional neural networks
• Cosmic nucleosynthesis: a multi-messenger challenge
• The stellar parameters and elemental abundances from low-resolution spectra I: 1.2 million giants from LAMOST DR8
• Combined APOGEE-GALAH stellar catalogues using the Cannon
• Assessing the performance of LTE and NLTE synthetic stellar spectra in a machine learning framework
• Stellar Populations With Optical Spectra: Deep Learning vs. Popular Spectrum Fitting Codes
• Estimating Stellar Parameters from LAMOST Low-resolution Spectra
• Titans metal-poor reference stars. I. Accurate effective temperatures and surface gravities for dwarfs and subgiants from 3D non-LTE H$\alpha$ profiles and Gaia parallaxes
• Analysis of Stellar Spectra from LAMOST DR5 with Generative Spectrum Networks
• Emission-line diagnostics of HII regions using conditional Invertible Neural Networks
• Deep Learning in Searching the Spectroscopic Redshift of Quasars
• The Gaia-ESO Survey: Chemical evolution of Mg and Al in the Milky Way with Machine-Learning
• The Gaia-ESO Survey: Preparing the ground for 4MOST & WEAVE galactic surveys. Chemical evolution of lithium with machine learning
• Machine Learning Approach to Integral Field Unit Spectroscopy Observations: I. HII Region Kinematics
• Using autoencoders and deep transfer learning to determine the stellar parameters of 286 CARMENES M dwarfs
• Constraining stellar population parameters from narrow band photometric surveys using convolutional neural networks
• Artificial Intelligence Assisted Inversion (AIAI) of Synthetic Type Ia Supernova Spectra
• The Swan: Data-Driven Inference of Stellar Surface Gravities for Cool Stars from Photometric Light Curves
• A Model RRNet for Spectral Information Exploitation and LAMOST Medium-resolution Spectrum Parameter Estimation
• A Machine Learning Approach to Integral Field Unit Spectroscopy Observations: II. HII Region LineRatios
• Galaxy Spectra neural Network (GaSNet). II. Using Deep Learning for Spectral Classification and Redshift Predictions
• SpectroTranslator: a deep-neural network algorithm to homogenize spectroscopic parameters
• The CARMENES search for exoplanets around M dwarfs -- A deep transfer learning method to determine Teff and [M/H] of target stars
• Reliable stellar abundances of individual stars with the MUSE integral-field spectrograph
• Spectral classification of young stars using conditional invertible neural networks I. Introducing and validating the method
• Galaxy Spectra Networks (GaSNet). III. Generative pre-trained network for spectrum reconstruction, redshift estimate and anomaly detection
• Parameter Estimation of LAMOST Medium-resolution Stellar Spectra
• Stellar Cluster Detection using GMM with Deep Variational Autoencoder
• The GALAH survey: Improving chemical abundances using star clusters
• StarUnLink: identifying and mitigating signals from communications satellites in stellar spectral surveys
• SpecPT (Spectroscopy Pre-trained Transformer) Model for Extragalactic Spectroscopy: I. Architecture and Automated Redshift Measurement
• CLOVER: Convnet Line-fitting Of Velocities in Emission-line Regions
• A Machine Learning Approach to Integral Field Unit Spectroscopy Observations: III. Disentangling Multiple Components in Hii regions
• Estimating Atmospheric Parameters of DA White Dwarf Stars with Deep Learning
• The Survey of Surveys: machine learning for stellar parametrization
• Survey of Surveys. II. Stellar parameters for 23 millions of stars
• Spectral classification of young stars using conditional invertible neural networks II. Application to Trumpler 14 in Carina
• PSF-based Analysis for Detecting Unresolved Wide Binaries
• Nuclear astrophysics
• LRPayne: Stellar parameters and abundances from low-resolution spectra
• SPAN: A cross-platform Python GUI software for optical and near-infrared spectral analysis
• Estimating stellar atmospheric parameters and elemental abundances using fully connected residual network
• Metallicities in M dwarfs: Investigating different determination techniques
• Playing CHESS with stars. I. Search for similar stars in large spectroscopic data sets