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

TL;DR

QuasarNET is a deep neural network that achieves human-level accuracy in classifying and estimating redshifts of astrophysical spectra, significantly reducing errors and confusion in spectral feature detection.

Contribution

It introduces a feature detection approach for spectral classification and redshift estimation using deep learning, outperforming traditional methods in accuracy and reliability.

Findings

Achieves 99.51% purity and 99.52% completeness in quasar identification.

Reduces catastrophic redshift failures to below 0.2%.

Classifies BAL features with over 97% accuracy.

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 $99.51\pm0.03$\% pure and $99.52\pm0.03$\% 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 $98.0\pm0.4$\% for recognizing BAL and $97.0\pm0.2$\% 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.