Harvesting the Lyα forest with convolutional neural networks

Abstract

We develop a machine learning based algorithm using a convolutional neural network (CNN) to identify low HI column density Ly$\alpha$ absorption systems ($\log N_{\mathrm{HI}}/{\mathrm{cm}}^{-2}<17$) in the Ly$\alpha$ forest, and predict their physical properties, such as their HI column density ($\log N_{\mathrm{HI}}/{\mathrm{cm}}^{-2}$), redshift ($z_{\mathrm{HI}}$), and Doppler width ($b_{\mathrm{HI}}$). Our CNN models are trained using simulated spectra (S/N $\simeq 10$), and we test their performance on high quality spectra of quasars at redshift $z\sim 2.5-2.9$ observed with the High Resolution Echelle Spectrometer on the Keck I telescope. We find that $\sim 78\%$ of the systems identified by our algorithm are listed in the manual Voigt profile fitting catalogue. We demonstrate that the performance of our CNN is stable and consistent for all simulated and observed spectra with S/N $\gtrsim 10$. Our model can therefore be consistently used to analyse the enormous number of both low and high S/N data available with current and future facilities. Our CNN provides state-of-the-art predictions within the range $12.5\le \log N_{\mathrm{HI}}/\mathrm{c}{\mathrm{m}}^{-2}<15.5$ with a mean absolute error of $\Delta (\log N_{\mathrm{HI}}/{\mathrm{cm}}^{-2})=0.13$, $\Delta (z_{\mathrm{HI}})=2.7\times {10}^{-5}$, and $\Delta (b_{\mathrm{HI}})=4.1\mathrm{km}{\mathrm{s}}^{-1}$. The CNN prediction costs $<3$ minutes per model per spectrum with a size of 120\,000 pixels using a laptop computer. We demonstrate that CNNs can significantly increase the efficiency of analysing Ly$\alpha$ forest spectra, and thereby greatly increase the statistics of Ly$\alpha$ absorbers.

22 pages including 2-pages Appendices, 14 figures plus 4 figures in Appendices. This paper is submitted to MNRAS and has addressed the first referee report

References in corpus (14)

• An Overview of Multi-Task Learning in Deep Neural Networks
• On the (Non)Evolution of HI Disks over Cosmic Time
• 4MOST: Project overview and information for the First Call for Proposals
• Possible evidence for an inverted temperature-density relation in the intergalactic medium from the flux distribution of the Lyman-alpha forest
• Physical Properties, Baryon Content, and Evolution of the LyαForest: New Insights from High Resolution Observations at z < 0.4
• Deep learning predictions of galaxy merger stage and the importance of observational realism
• Galaxy Merger Rates up to z $\sim$ 3 using a Bayesian Deep Learning Model $-$ A Major-Merger classifier using IllustrisTNG Simulation data
• The Second Data Release of the KODIAQ Survey
• Photo-heating and the fate of hard photons during the reionisation of HeII by quasars
• Galaxy Morphological Classification Catalogue of the Dark Energy Survey Year 3 data with Convolutional Neural Networks
• VoIgt profile Parameter Estimation Routine (VIPER): H I photoionization rate at z<0.5
• WEAVE-QSO: A Massive Intergalactic Medium Survey for the William Herschel Telescope
• The evolution of the low-density HI intergalactic medium from z=3.6 to 0: Data, transmitted flux and HI column density
• Deep Learning of DESI Mock Spectra to Find Damped Lyα Systems