Automated quasar continuum estimation using neural networks: a comparative study of deep-learning architectures

Abstract

Context. Ongoing and upcoming large spectroscopic surveys are drastically increasing the number of observed quasar spectra, requiring the development of fast and accurate automated methods to estimate spectral continua. Aims. This study evaluates the performance of three neural networks (NN) - an autoencoder, a convolutional NN (CNN), and a U-Net - in predicting quasar continua within the rest-frame wavelength range of $1020\text{A˚}$ to $2000\text{A˚}$. The ability to generalize and predict galaxy continua within the range of $3500\text{A˚}$ to $5500\text{A˚}$ is also tested. Methods. The performance of these architectures is evaluated using the absolute fractional flux error (AFFE) on a library of mock quasar spectra for the WEAVE survey, and on real data from the Early Data Release observations of the Dark Energy Spectroscopic Instrument (DESI) and the VIMOS Public Extragalactic Redshift Survey (VIPERS). Results. The autoencoder outperforms the U-Net, achieving a median AFFE of 0.009 for quasars. The best model also effectively recovers the Ly$\alpha$ optical depth evolution in DESI quasar spectra. With minimal optimization, the same architectures can be generalized to the galaxy case, with the autoencoder reaching a median AFFE of 0.014 and reproducing the D4000n break in DESI and VIPERS galaxies.

23 pages, 16 figures. Accepted in A&A

References in corpus (69)

• The Seventh Data Release of the Sloan Digital Sky Survey
• Road Extraction by Deep Residual U-Net
• Evolutionary population synthesis: models, analysis of the ingredients and application to high-z galaxies
• Improving the full spectrum fitting method: accurate convolution with Gauss-Hermite functions
• A fundamental relation between mass, SFR and metallicity in local and high redshift galaxies
• The Sloan Digital Sky Survey Quasar Catalog: Sixteenth Data Release
• The Baryon Census in a Multiphase Intergalactic Medium: 30% of the Baryons May Still Be Missing
• The Sloan Digital Sky Survey Quasar Catalog: Fourteenth Data Release
• The Sloan Digital Sky Survey Quasar Catalog: twelfth data release
• The Mass-Metallicity and the Fundamental Metallicity Relation revisited on a fully Te-based abundance scale for galaxies
• Stellar population models at high spectral resolution
• The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: Baryon acoustic oscillations with Lyman-$\alpha$ forests
• The VIMOS Public Extragalactic Redshift Survey (VIPERS). An unprecedented view of galaxies and large-scale structure at 0.5<z<1.2
• The Early Data Release of the Dark Energy Spectroscopic Instrument
• The Universal Relation of Galactic Chemical Evolution: The Origin of the Mass-Metallicity Relation
• Physical Conditions in a Young, Unreddened, Low Metallicity Galaxy at High Redshift
• Full spectrum fitting with photometry in pPXF: stellar population versus dynamical masses, non-parametric star formation history and metallicity for 3200 LEGA-C galaxies at redshift z~0.8
• The Sloan Digital Sky Survey quasar catalog: ninth data release
• The VIMOS Public Extragalactic Survey (VIPERS): First Data Release of 57 204 spectroscopic measurements
• The Sloan Digital Sky Survey quasar catalog: tenth data release
• A Direct Precision Measurement of the Intergalactic Lyman-alpha Opacity at 2<z<4.2
• The VIMOS Public Extragalactic Redshift Survey (VIPERS). Full spectroscopic data and auxiliary information release (PDR-2)
• The wide-field, multiplexed, spectroscopic facility WEAVE: Survey design, overview, and simulated implementation
• A Refined Measurement of the Mean Transmitted Flux in the Ly-alpha Forest over 2 < z < 5 Using Composite Quasar Spectra
• Photometric redshift analysis in the Dark Energy Survey Science Verification data
• The Sherwood simulation suite: overview and data comparisons with the Lyman-alpha forest at redshifts 2 < z < 5
• A Principal Component Analysis of quasar UV spectra at z~3
• Lyα emission line reconstruction for high-$z$ QSOs
• Fundamental metallicity relation in CALIFA, SDSS-IV MaNGA and high-z galaxies
• MUSE Analysis of Gas around Galaxies (MAGG) -- II: Metal-enriched halo gas around z~1 galaxies
• The H I opacity of the intergalactic medium at redshifts 1.6 < z < 3.2
• Unsupervised self-organised mapping: a versatile empirical tool for object selection, classification and redshift estimation in large surveys
• Dimensionality Reduction of SDSS Spectra with Variational Autoencoders
• Predicting Quasar Continua Near Lyman-$\alpha$ with Principal Component Analysis
• Constraints on the low-mass end of the mass-metallicity relation at z=1-2 from lensed galaxies
• Mean-flux Regulated PCA Continuum Fitting of SDSS Lyman-alpha Forest Spectra
• A Survey of MgII Absorption at 2 < z < 6 with Magellan / FIRE: I: Sample and Evolution of the MgII Frequency
• Catalogues of Active Galactic Nuclei From Gaia and unWISE Data
• Precious Metals in SDSS Quasar Spectra I: Tracking the Evolution of Strong, 1.5 < z < 4.5 CIV Absorbers with Thousands of Systems
• Mg II Absorption at 2<z<7 with Magellan/FIRE, III. Full Statistics of Absorption Towards 100 High-Redshift QSOs
• Machine learning technique for morphological classification of galaxies from the SDSS. I. Photometry-based approach
• Data augmentation for machine learning redshifts applied to SDSS galaxies
• The XQR-30 Metal Absorber Catalog: 778 Absorption Systems Spanning 2 < z < 6.5
• The evolution of the Si IV content in the Universe from the epoch of reionization to cosmic noon
• The VIMOS Public Extragalactic Redshift Survey (VIPERS). AGN feedback in [NeV] emitters
• MUSE Analysis of Gas around Galaxies (MAGG) -- V: Linking ionized gas traced by CIV and SiIV absorbers to Ly$\alpha$ emitting galaxies at $z\approx 3.0-4.5$
• Autoencoding Galaxy Spectra I: Architecture
• Stacking for machine learning redshifts applied to SDSS galaxies
• Examining the Decline in the C IV Content of the Universe over 4.3 < z < 6.3 using the E-XQR-30 Sample
• The Type II AGN-host galaxy connection: insights from the VVDS and VIPERS surveys
• SFR estimations from z=0 to z=0.9 -- A comparison of SFR calibrators for star-forming galaxies
• Autoencoding Galaxy Spectra II: Redshift Invariance and Outlier Detection
• Inferring physical properties of galaxies from their emission line spectra
• GAME: GAlaxy Machine learning for Emission lines
• A Deep Learning Approach to Quasar Continuum Prediction
• Mapping the Diversity of Galaxy Spectra with Deep Unsupervised Machine Learning
• Strong Mg II and Fe II Absorbers at 2.2 < z < 6.0
• Mass-Metallicity Relation for Local Analogs of High-Redshift galaxies: Implications for the Evolution of the Mass-Metallicity Relations
• WEAVE-StePS. A stellar population survey using WEAVE at WHT
• Quasar Factor Analysis -- An Unsupervised and Probabilistic Quasar Continuum Prediction Algorithm with Latent Factor Analysis
• Darth Fader: Using wavelets to obtain accurate redshifts of spectra at very low signal-to-noise
• The Gaia-ESO Survey: Chemical evolution of Mg and Al in the Milky Way with Machine-Learning
• Measurements of effective optical depth in the Ly$\alpha$ forest from the BOSS DR12 Quasar sample
• Constraining stellar population parameters from narrow band photometric surveys using convolutional neural networks
• The interstellar medium of dwarf galaxies: new insights from Machine Learning analysis of emission line spectra
• The fundamental metallicity relation from SDSS (z~0) to VIPERS (z~0.7) Data selection or evolution
• Machine learning the gap between real and simulated nebulae: A domain-adaptation approach to classify ionised nebulae in nearby galaxies
• Inferring redshift and galaxy properties via a multi-task neural net with probabilistic outputs: An application to simulated MOONS spectra
• Identifying reionization-epoch galaxies with extreme levels of Lyman continuum leakage in James Webb Space Telescope surveys

Cited by in corpus (1)

• QUEST (Quasar Unsupervised Encoder and Synthesis Tool): A machine learning framework to generate quasar spectra