Paper detail

Deep Learning in Searching the Spectroscopic Redshift of Quasars

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 &#34;\textit{local}&#34; and &#34;\textit{global}&#34; 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, $|Δν|< 6000~ \rm km/s$ and 98.0$\%$ for $|Δν|< 12000~ \rm km/s$. 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 $|Δν|< 6000~ \rm km/s$ and 97.9$\%$ for $|Δν|< 12000~ \rm km/s$. 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.