Application of convolutional neural networks for stellar spectral classification
Abstract
ABSTRACT Due to the ever-expanding volume of observed spectroscopic data from surveys such as SDSS and LAMOST, it has become important to apply artificial intelligence (AI) techniques for analysing stellar spectra to solve spectral classification and regression problems like the determination of stellar atmospheric parameters Teff, $$\rm {\log g}$$, and [Fe/H]. We propose an automated approach for the classification of stellar spectra in the optical region using convolutional neural networks (CNNs). Traditional machine learning (ML) methods with ‘shallow’ architecture (usually up to two hidden layers) have been trained for these purposes in the past. However, deep learning methods with a larger number of hidden layers allow the use of finer details in the spectrum which results in improved accuracy and better generalization. Studying finer spectral signatures also enables us to determine accurate differential stellar parameters and find rare objects. We examine various machine and deep learning algorithms like artificial neural networks, Random Forest, and CNN to classify stellar spectra using the Jacoby Atlas, ELODIE, and MILES spectral libraries as training samples. We test the performance of the trained networks on the Indo-U.S. Library of Coudé Feed Stellar Spectra (CFLIB). We show that using CNNs, we are able to lower the error up to 1.23 spectral subclasses as compared to that of two subclasses achieved in the past studies with ML approach. We further apply the trained model to classify stellar spectra retrieved from the SDSS data base with SNR > 20.
- Authors:
-
- Inter University Centre for Astronomy and Astrophysics, Pune 411007, India
- PRIOR, Allen Institute for Artificial Intelligence, Seattle 98103, USA, Department of Computer Science and Engineering, University of Washington, Seattle 98195, USA
- Indian Institute of Astrophysics, Koramangala, Bengaluru 560034, India
- Department of Physics, Mar Thoma College, Chungathara 679334, Nilambur, India
- Publication Date:
- Sponsoring Org.:
- USDOE
- OSTI Identifier:
- 1706198
- Resource Type:
- Published Article
- Journal Name:
- Monthly Notices of the Royal Astronomical Society
- Additional Journal Information:
- Journal Name: Monthly Notices of the Royal Astronomical Society Journal Volume: 491 Journal Issue: 2; Journal ID: ISSN 0035-8711
- Publisher:
- Oxford University Press
- Country of Publication:
- United Kingdom
- Language:
- English
Citation Formats
Sharma, Kaushal, Kembhavi, Ajit, Kembhavi, Aniruddha, Sivarani, T., Abraham, Sheelu, and Vaghmare, Kaustubh. Application of convolutional neural networks for stellar spectral classification. United Kingdom: N. p., 2019.
Web. doi:10.1093/mnras/stz3100.
Sharma, Kaushal, Kembhavi, Ajit, Kembhavi, Aniruddha, Sivarani, T., Abraham, Sheelu, & Vaghmare, Kaustubh. Application of convolutional neural networks for stellar spectral classification. United Kingdom. https://doi.org/10.1093/mnras/stz3100
Sharma, Kaushal, Kembhavi, Ajit, Kembhavi, Aniruddha, Sivarani, T., Abraham, Sheelu, and Vaghmare, Kaustubh. Wed .
"Application of convolutional neural networks for stellar spectral classification". United Kingdom. https://doi.org/10.1093/mnras/stz3100.
@article{osti_1706198,
title = {Application of convolutional neural networks for stellar spectral classification},
author = {Sharma, Kaushal and Kembhavi, Ajit and Kembhavi, Aniruddha and Sivarani, T. and Abraham, Sheelu and Vaghmare, Kaustubh},
abstractNote = {ABSTRACT Due to the ever-expanding volume of observed spectroscopic data from surveys such as SDSS and LAMOST, it has become important to apply artificial intelligence (AI) techniques for analysing stellar spectra to solve spectral classification and regression problems like the determination of stellar atmospheric parameters Teff, $\rm {\log g}$, and [Fe/H]. We propose an automated approach for the classification of stellar spectra in the optical region using convolutional neural networks (CNNs). Traditional machine learning (ML) methods with ‘shallow’ architecture (usually up to two hidden layers) have been trained for these purposes in the past. However, deep learning methods with a larger number of hidden layers allow the use of finer details in the spectrum which results in improved accuracy and better generalization. Studying finer spectral signatures also enables us to determine accurate differential stellar parameters and find rare objects. We examine various machine and deep learning algorithms like artificial neural networks, Random Forest, and CNN to classify stellar spectra using the Jacoby Atlas, ELODIE, and MILES spectral libraries as training samples. We test the performance of the trained networks on the Indo-U.S. Library of Coudé Feed Stellar Spectra (CFLIB). We show that using CNNs, we are able to lower the error up to 1.23 spectral subclasses as compared to that of two subclasses achieved in the past studies with ML approach. We further apply the trained model to classify stellar spectra retrieved from the SDSS data base with SNR > 20.},
doi = {10.1093/mnras/stz3100},
journal = {Monthly Notices of the Royal Astronomical Society},
number = 2,
volume = 491,
place = {United Kingdom},
year = {Wed Nov 06 00:00:00 EST 2019},
month = {Wed Nov 06 00:00:00 EST 2019}
}
https://doi.org/10.1093/mnras/stz3100
Works referenced in this record:
The Baryon Oscillation Spectroscopic Survey of Sdss-Iii
journal, December 2012
- Dawson, Kyle S.; Schlegel, David J.; Ahn, Christopher P.
- The Astronomical Journal, Vol. 145, Issue 1
Deep-learnt classification of light curves
conference, November 2017
- Mahabal, A.; Sheth, K.; Gieseke, F.
- 2017 IEEE Symposium Series on Computational Intelligence (SSCI)
A Triple Main Sequence in the Globular Cluster NGC 2808
journal, April 2007
- Piotto, G.; Bedin, L. R.; Anderson, J.
- The Astrophysical Journal, Vol. 661, Issue 1
A Stellar Spectral Flux Library: 1150–25000 Å
journal, July 1998
- Pickles, A. J.
- Publications of the Astronomical Society of the Pacific, Vol. 110, Issue 749
Measurement of Inequality of Incomes
journal, March 1921
- Gini, Corrado
- The Economic Journal, Vol. 31, Issue 121
Stellar spectral classification using automated schemes
journal, May 1994
- Gulati, R. K.; Gupta, Ranjan; Gothoskar, Pradeep
- The Astrophysical Journal, Vol. 426
Maps of Dust Infrared Emission for Use in Estimation of Reddening and Cosmic Microwave Background Radiation Foregrounds
journal, June 1998
- Schlegel, David J.; Finkbeiner, Douglas P.; Davis, Marc
- The Astrophysical Journal, Vol. 500, Issue 2
Rotation Velocities of Red and Blue Field Horizontal‐Branch Stars
journal, November 2003
- Behr, Bradford B.
- The Astrophysical Journal Supplement Series, Vol. 149, Issue 1
Kinematics of Metal-poor Stars in the Galaxy. II. Proper Motions for a Large Nonkinematically Selected Sample
journal, June 2000
- Beers, Timothy C.; Chiba, Masashi; Yoshii, Yuzuru
- The Astronomical Journal, Vol. 119, Issue 6
An Empirical Template Library of Stellar Spectra for a Wide Range of Spectral Classes, Luminosity Classes, and Metallicities Using SDSS BOSS Spectra
journal, June 2017
- Kesseli, Aurora Y.; West, Andrew A.; Veyette, Mark
- The Astrophysical Journal Supplement Series, Vol. 230, Issue 2
Searching for exoplanets using artificial intelligence
journal, October 2017
- Pearson, Kyle A.; Palafox, Leon; Griffith, Caitlin A.
- Monthly Notices of the Royal Astronomical Society, Vol. 474, Issue 1
An application of deep learning in the analysis of stellar spectra
journal, December 2017
- Fabbro, S.; Venn, K. A.; O'Briain, T.
- Monthly Notices of the Royal Astronomical Society, Vol. 475, Issue 3
A library of stellar spectra
journal, October 1984
- Jacoby, G. H.; Hunter, D. A.; Christian, C. A.
- The Astrophysical Journal Supplement Series, Vol. 56
Automated classification of stellar spectra -- II. Two-dimensional classification with neural networks and principal components analysis
journal, August 1998
- Bailer-Jones, C. A. L.; Irwin, M.; Von Hippel, T.
- Monthly Notices of the Royal Astronomical Society, Vol. 298, Issue 2
Abundances for metal-poor stars with accurate parallaxes: I. Basic data
journal, May 2003
- Gratton, R. G.; Carretta, E.; Claudi, R.
- Astronomy & Astrophysics, Vol. 404, Issue 1
A catalogue of spectral classification and photometric data of B-type stars
journal, January 1993
- Hamdy, M. A.; Abo Elazm, M. S.; Saad, S. M.
- Astrophysics and Space Science, Vol. 203, Issue 1
Coudé-feed stellar spectral library – atmospheric parameters
journal, December 2010
- Wu, Yue; Singh, H. P.; Prugniel, P.
- Astronomy & Astrophysics, Vol. 525
Updating quasi-Newton matrices with limited storage
journal, September 1980
- Nocedal, Jorge
- Mathematics of Computation, Vol. 35, Issue 151
Mixed Populations in Globular Clusters: et tu, 47 tuc?
journal, April 2009
- Anderson, J.; Piotto, G.; King, I. R.
- The Astrophysical Journal, Vol. 697, Issue 1
Automatic spectral classification of stellar spectra with low signal-to-noise ratio using artificial neural networks
journal, February 2012
- Navarro, S. G.; Corradi, R. L. M.; Mampaso, A.
- Astronomy & Astrophysics, Vol. 538
Spectral Classification and Redshift Measurement for the Sdss-Iii Baryon Oscillation Spectroscopic Survey
journal, October 2012
- Bolton, Adam S.; Schlegel, David J.; Aubourg, Éric
- The Astronomical Journal, Vol. 144, Issue 5
The Indo‐US Library of Coude Feed Stellar Spectra
journal, June 2004
- Valdes, Francisco; Gupta, Ranjan; Rose, James A.
- The Astrophysical Journal Supplement Series, Vol. 152, Issue 2
A spectroscopic survey for λ Bootis stars: II. The observational data
journal, July 2001
- Paunzen, E.; Duffee, B.; Heiter, U.
- Astronomy & Astrophysics, Vol. 373, Issue 2
Accurate Two‐Dimensional Classification of Stellar Spectra with Artificial Neural Networks
journal, October 1997
- Weaver, Wm. Bruce; Torres‐Dodgen, Ana V.
- The Astrophysical Journal, Vol. 487, Issue 2
The Kinematics of the Gould Belt: an Expanding Group?
journal, December 1968
- Rountree Lesh, Janet
- The Astrophysical Journal Supplement Series, Vol. 17
G. P. Kuiper's spectral classifications of proper-motion stars
journal, October 1985
- Bidelman, W. P.
- The Astrophysical Journal Supplement Series, Vol. 59
An absolutely calibrated T eff scale from the infrared flux method : Dwarfs and subgiants
journal, March 2010
- Casagrande, L.; Ramírez, I.; Meléndez, J.
- Astronomy and Astrophysics, Vol. 512
Reducing the Dimensionality of Data with Neural Networks
journal, July 2006
- Hinton, G. E.
- Science, Vol. 313, Issue 5786
Consistent distances from Baade-Wesselink analyses of Cepheids and RR Lyraes
journal, July 2003
- Kovacs, G.
- Monthly Notices of the Royal Astronomical Society, Vol. 342, Issue 4
Lithium in an extreme halo star
journal, March 1988
- Hobbs, L. M.; Pilachowski, Catherine
- The Astrophysical Journal, Vol. 326
An Analysis of the Shapes of Interstellar Extinction Curves. V. The IR‐through‐UV Curve Morphology
journal, July 2007
- Fitzpatrick, E. L.; Massa, D.
- The Astrophysical Journal, Vol. 663, Issue 1
Smoothing and Differentiation of Data by Simplified Least Squares Procedures.
journal, July 1964
- Savitzky, Abraham.; Golay, M. J. E.
- Analytical Chemistry, Vol. 36, Issue 8
Stellar spectral classification using principal component analysis and artificial neural networks
journal, April 1998
- Singh, H. P.; Gulati, R. K.; Gupta, R.
- Monthly Notices of the Royal Astronomical Society, Vol. 295, Issue 2
Rotation-invariant convolutional neural networks for galaxy morphology prediction
journal, April 2015
- Dieleman, Sander; Willett, Kyle W.; Dambre, Joni
- Monthly Notices of the Royal Astronomical Society, Vol. 450, Issue 2
Medium-resolution Isaac Newton Telescope library of empirical spectra
journal, September 2006
- Sanchez-Blazquez, P.; Peletier, R. F.; Jimenez-Vicente, J.
- Monthly Notices of the Royal Astronomical Society, Vol. 371, Issue 2
Spectral classification of high-proper-motion stars
journal, May 1984
- Lee, S. -G.
- The Astronomical Journal, Vol. 89
Empirical calibration of the λ 4000 Å
break
journal, October 1999
- Gorgas, J.; Cardiel, N.; Pedraz, S.
- Astronomy and Astrophysics Supplement Series, Vol. 139, Issue 1
An Expert Computer Program for Classifying Stars on the mk Spectral Classification System
journal, March 2014
- Gray, R. O.; Corbally, C. J.
- The Astronomical Journal, Vol. 147, Issue 4
Gaia FGK benchmark stars: Effective temperatures and surface gravities
journal, October 2015
- Heiter, U.; Jofré, P.; Gustafsson, B.
- Astronomy & Astrophysics, Vol. 582
The atmospheric parameters and spectral interpolator for the MILES stars
journal, July 2011
- Prugniel, Ph.; Vauglin, I.; Koleva, M.
- Astronomy & Astrophysics, Vol. 531
The Oosterhoff period-metallicity relation for RR Lyrae stars at the blue fundamental edge of the instability strip.
journal, August 1993
- Sandage, Allan
- The Astronomical Journal, Vol. 106
Detection of bars in galaxies using a deep convolutional neural network
journal, March 2018
- Abraham, Sheelu; Aniyan, A. K.; Kembhavi, Ajit K.
- Monthly Notices of the Royal Astronomical Society, Vol. 477, Issue 1
New atmospheric parameters and spectral interpolator for the MILES cool stars
journal, December 2015
- Sharma, Kaushal; Prugniel, Philippe; Singh, Harinder P.
- Astronomy & Astrophysics, Vol. 585
Scalable training of L 1 -regularized log-linear models
conference, January 2007
- Andrew, Galen; Gao, Jianfeng
- Proceedings of the 24th international conference on Machine learning - ICML '07
Learning Deep Architectures for AI
journal, January 2009
- Bengio, Y.
- Foundations and Trends® in Machine Learning, Vol. 2, Issue 1
A database of high and medium-resolution stellar spectra
journal, April 2001
- Prugniel, Ph.; Soubiran, C.
- Astronomy & Astrophysics, Vol. 369, Issue 3
Starmind: a Fuzzy Logic Knowledge-Based System for the Automated Classification of Stars in the mk System
journal, January 2009
- Manteiga, M.; Carricajo, I.; Rodríguez, A.
- The Astronomical Journal, Vol. 137, Issue 2
The luminosity scale of RR Lyrae stars with the Baade-Wesselink method. II - The absolute magnitudes of 13 field RR Lyrae stars
journal, May 1990
- Liu, Tianxing; Janes, Kenneth A.
- The Astrophysical Journal, Vol. 354
Neutron‐Capture Elements in the Early Galaxy: Insights from a Large Sample of Metal‐poor Giants
journal, November 2000
- Burris, Debra L.; Pilachowski, Catherine A.; Armandroff, Taft E.
- The Astrophysical Journal, Vol. 544, Issue 1
Extracting and composing robust features with denoising autoencoders
conference, January 2008
- Vincent, Pascal; Larochelle, Hugo; Bengio, Yoshua
- Proceedings of the 25th international conference on Machine learning - ICML '08
Carbon and oxygen abundances in stellar populations
journal, August 2014
- Nissen, P. E.; Chen, Y. Q.; Carigi, L.
- Astronomy & Astrophysics, Vol. 568
Non-LTE line formation for heavy elements in four very metal-poor stars
journal, December 2007
- Mashonkina, L.; Zhao, G.; Gehren, T.
- Astronomy & Astrophysics, Vol. 478, Issue 2
Discoveries on Southern Objective-Prism Plates. II. New Probable Field Horizontal-Branch Stars
journal, February 1971
- MacConnell, D. J.; Frye, R. L.; Bidelman, W. P.
- Publications of the Astronomical Society of the Pacific, Vol. 83
PEPSI deep spectra: II. Gaia benchmark stars and other M-K standards
journal, April 2018
- Strassmeier, K. G.; Ilyin, I.; Weber, M.
- Astronomy & Astrophysics, Vol. 612
Measuring Reddening with Sloan Digital sky Survey Stellar Spectra and Recalibrating sfd
journal, August 2011
- Schlafly, Edward F.; Finkbeiner, Douglas P.
- The Astrophysical Journal, Vol. 737, Issue 2
Contributions to the Nearby Stars (NStars) Project: Spectroscopy of Stars Earlier than M0 within 40 Parsecs: The Northern Sample. I.
journal, October 2003
- Gray, R. O.; Corbally, C. J.; Garrison, R. F.
- The Astronomical Journal, Vol. 126, Issue 4
S$\mathsf{_{B^9}}$: The ninth catalogue of spectroscopic binary orbits
journal, August 2004
- Pourbaix, D.; Tokovinin, A. A.; Batten, A. H.
- Astronomy & Astrophysics, Vol. 424, Issue 2
Abundances in the Local Region. ii. f, g, and k Dwarfs and Subgiants
journal, December 2016
- Luck, R. Earle
- The Astronomical Journal, Vol. 153, Issue 1
Multiple Stellar Populations in the Galactic Globular Cluster ngc 6752
journal, January 2010
- Milone, A. P.; Piotto, G.; King, I. R.
- The Astrophysical Journal, Vol. 709, Issue 2
The Gaia FGK Benchmark Stars Version 2.1
journal, July 2018
- Jofré, Paula; Heiter, Ulrike; Maia, Marcelo Tucci
- Research Notes of the AAS, Vol. 2, Issue 3
Radio Galaxy Zoo: Claran – a deep learning classifier for radio morphologies
journal, October 2018
- Wu, Chen; Wong, Oiwei Ivy; Rudnick, Lawrence
- Monthly Notices of the Royal Astronomical Society, Vol. 482, Issue 1
MK classifications for F-and G-type stars. II.
journal, March 1970
- Harlan, E. A.; Taylor, D. C.
- The Astronomical Journal, Vol. 75
Analysis of neutron capture elements in metal-poor stars
journal, May 2001
- Mishenina, T. V.; Kovtyukh, V. V.
- Astronomy & Astrophysics, Vol. 370, Issue 3
The 13th Data Release of the Sloan Digital Sky Survey: First Spectroscopic Data from the SDSS-IV Survey Mapping Nearby Galaxies at Apache Point Observatory
journal, December 2017
- Albareti, Franco D.; Prieto, Carlos Allende; Almeida, Andres
- The Astrophysical Journal Supplement Series, Vol. 233, Issue 2