Deep Learning in Searching the Spectroscopic Redshift of Quasars

TL;DR

This paper introduces FNet, a deep residual CNN that accurately estimates quasar redshifts from SDSS spectra, outperforming traditional methods especially on spectra lacking clear emission lines, thus enhancing automated redshift determination.

Contribution

The paper presents FNet, a novel deep residual CNN architecture with 24 layers that improves redshift estimation accuracy and applicability over existing methods like QuasarNET.

Findings

FNet achieves 97.0% accuracy for |Δν|<6000 km/s.

FNet outperforms QuasarNET on spectra with missing emission lines.

FNet is suitable for a wider range of SDSS spectra, reducing manual inspection.

Abstract

Studying the cosmological sources at their cosmological rest-frames is crucial to track the cosmic history and properties of compact objects. In view of the increasing data volume of existing and upcoming telescopes/detectors, we here construct a 1--dimensional convolutional neural network (CNN) with a residual neural network (ResNet) structure to estimate the redshift of quasars in Sloan Digital Sky Survey IV (SDSS-IV) catalog from DR16 quasar-only (DR16Q) of eBOSS on a broad range of signal-to-noise ratios, named \code{FNet}. Owing to its $24$ convolutional layers and the ResNet structure with different kernel sizes of $500$, $200$ and $15$, FNet is able to discover the "\textit{local}" and "\textit{global}" patterns in the whole sample of spectra by a self-learning procedure. It reaches the accuracy of 97.0$\%$ for the velocity difference for redshift, $|\Delta\nu|< 6000~ \rm km/s$ and 98.0$\%$ for $|\Delta\nu|< 12000~ \rm km/s$. While \code{QuasarNET}, which is a standard CNN adopted in the SDSS routine and is constructed by 4 convolutional layers (no ResNet structure), with kernel sizes of $10$, to measure the redshift via identifying seven emission lines (\textit{local} patterns), fails in estimating redshift of $\sim 1.3\%$ of visually inspected quasars in DR16Q catalog, and it gives 97.8$\%$ for $|\Delta\nu|< 6000~ \rm km/s$ and 97.9$\%$ for $|\Delta\nu|< 12000~ \rm km/s$. Hence, FNet provides similar accuracy to \code{QuasarNET}, but it is applicable for a wider range of SDSS spectra, especially for those missing the clear emission lines exploited by \code{QuasarNET}. These properties of \code{FNet}, together with the fast predictive power of machine learning, allow \code{FNet} to be a more accurate alternative for the pipeline redshift estimator and can make it practical in the upcoming catalogs to reduce the number of spectra to visually inspect.