QuasarNET: Human-level spectral classification and redshifting with Deep Neural Networks

Abstract

We introduce QuasarNET, a deep convolutional neural network that performs classification and redshift estimation of astrophysical spectra with human-expert accuracy. We pose these two tasks as a \emph{feature detection} problem: presence or absence of spectral features determines the class, and their wavelength determines the redshift, very much like human-experts proceed. When ran on BOSS data to identify quasars through their emission lines, QuasarNET defines a sample $99.51\pm 0.03$\% pure and $99.52\pm 0.03$\% complete, well above the requirements of many analyses using these data. QuasarNET significantly reduces the problem of line-confusion that induces catastrophic redshift failures to below 0.2\%. We also extend QuasarNET to classify spectra with broad absorption line (BAL) features, achieving an accuracy of $98.0\pm 0.4$\% for recognizing BAL and $97.0\pm 0.2$\% for rejecting non-BAL quasars. QuasarNET is trained on data of low signal-to-noise and medium resolution, typical of current and future astrophysical surveys, and could be easily applied to classify spectra from current and upcoming surveys such as eBOSS, DESI and 4MOST.

Submitted to MMRAS

Cited by in corpus (21)

• DESI 2024 IV: Baryon Acoustic Oscillations from the Lyman Alpha Forest
• The Spectroscopic Data Processing Pipeline for the Dark Energy Spectroscopic Instrument
• Target Selection and Validation of DESI Quasars
• The Million Quasars (Milliquas) Catalogue, v8
• The DESI Survey Validation: Results from Visual Inspection of the Quasar Survey Spectra
• Performance of the Quasar Spectral Templates for the Dark Energy Spectroscopic Instrument
• The Dark Energy Spectroscopic Instrument: One-dimensional power spectrum from first Lyman-$\alpha$ forest samples with Fast Fourier Transform
• The final SDSS-IV/SPIDERS X-ray point source spectroscopic catalogue
• Optimal strategies for identifying quasars in DESI
• Constraining primordial non-Gaussianity from DESI quasar targets and Planck CMB lensing
• Deep Learning in Searching the Spectroscopic Redshift of Quasars
• An extended catalog of galaxy-galaxy strong gravitational lenses discovered in DES using convolutional neural networks
• Spectroscopic QUasar Extractor and redshift (z) EstimatorSQUEzE I: Methodology
• Survey Operations for the Dark Energy Spectroscopic Instrument
• The miniJPAS survey quasar selection II: Machine learning classification with photometric measurements and uncertainties
• High redshift LBGs from deep broadband imaging for future spectroscopic surveys
• Identification confusion and blending concealment in the SDSS-DR16 Quasar catalogues -- 40 new quasars and 82 false quasars identified
• Detecting neutral hydrogen at z > 3 in large spectroscopic surveys of quasars
• Galaxy Spectra Networks (GaSNet). III. Generative pre-trained network for spectrum reconstruction, redshift estimate and anomaly detection
• Spectroscopic QUasar Extractor and redshift (z) EstimatorSQUEzE II: Universality of the results
• 3D Correlations in the Lyman-$\alpha$ Forest from Early DESI Data