• Corpus ID: 52133926

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

@article{Busca2018QuasarNETHS,
  title={QuasarNET: Human-level spectral classification and redshifting with Deep Neural Networks},
  author={Nicolas Busca and Christophe Balland},
  journal={ArXiv},
  year={2018},
  volume={abs/1808.09955}
}
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… 

Figures from this paper

Classification of Broad Absorption Line Quasars with a Convolutional Neural Network

Quasars that exhibit blueshifted, broad absorption lines (BAL QSOs) are an important probe of black hole feedback on galaxy evolution. Yet the presence of BALs is also a complication for large

Spectroscopic QUasar Extractor and redshift (z) Estimator squeze – I. Methodology

We present SQUEzE, a software package to classify quasar spectra and estimate their redshifts. SQUEzE is a random forest classifier operating on the parameters of candidate emission peaks identified

Spectroscopic QUasar extractor and redshift (z) estimator squeze – II. Universality of the results

In this paper we study the universality of the results of SQUEzE, a software package to classify quasar spectra and estimate their redshifts. The code is presented in \cite{Perez-Rafols+2019}. We

Detecting neutral hydrogen at z ≳ 3 in large spectroscopic surveys of quasars

We present a pipeline based on a random forest classifier for the identification of high column-density clouds of neutral hydrogen (i.e. the Lyman limit systems, LLSs) in absorption within large

Optimal strategies for identifying quasars in DESI

TLDR
This work demonstrates that the machine learning-based classifier QuasarNET will be of use for future surveys such as the Dark Energy Spectroscopic Instrument (DESI), comparing its performance to the DESI pipeline classifier redrock.

An Extended Catalog of Galaxy–Galaxy Strong Gravitational Lenses Discovered in DES Using Convolutional Neural Networks

We search Dark Energy Survey (DES) Year 3 imaging for galaxy–galaxy strong gravitational lenses using convolutional neural networks, extending previous work with new training sets and covering a

The Completed SDSS-IV Extended Baryon Oscillation Spectroscopic Survey: Baryon Acoustic Oscillations with Lyα Forests

We present a measurement of baryonic acoustic oscillations (BAOs) from Lyα absorption and quasars at an effective redshift using the complete extended Baryonic Oscillation Spectroscopic Survey

The final SDSS-IV/SPIDERS X-ray point source spectroscopic catalogue

Aims. We look to provide a detailed description of the SPectroscopic IDentification of ERosita Sources (SPIDERS) survey, an SDSS-IV programme aimed at obtaining spectroscopic classification and

Baryon acoustic oscillations from the cross-correlation of Lyα absorption and quasars in eBOSS DR14

We present a measurement of the baryon acoustic oscillation (BAO) scale at redshift z = 2.35 from the three-dimensional correlation of Lyman-α (Lyα) forest absorption and quasars. The study uses 266

Baryon acoustic oscillations at z = 2.34 from the correlations of Lyα absorption in eBOSS DR14

We measure the imprint of primordial baryon acoustic oscillations (BAOs) in the correlation function of Lyα absorption in quasar spectra from the Baryon Oscillation Spectroscopic Survey (BOSS) and

References

SHOWING 1-4 OF 4 REFERENCES

The DESI Experiment Part I: Science,Targeting, and Survey Design

DESI (Dark Energy Spectroscopic Instrument) is a Stage IV ground-based dark energy experiment that will study baryon acoustic oscillations (BAO) and the growth of structure through redshift-space

The DESI Experiment Part II: Instrument Design

DESI (Dark Energy Spectropic Instrument) is a Stage IV ground-based dark energy experiment that will study baryon acoustic oscillations and the growth of structure through redshift-space distortions

YOLO9000: Better, Faster, Stronger

TLDR
YOLO9000, a state-of-the-art, real-time object detection system that can detect over 9000 object categories, is introduced and a method to jointly train on object detection and classification is proposed, both novel and drawn from prior work.

Adam: A Method for Stochastic Optimization

TLDR
This work introduces Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments, and provides a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework.