Harvesting the Lyα forest with convolutional neural networks

Cheng, Ting-Yun; Cooke, Ryan J; Rudie, Gwen

Abstract

We develop a machine learning based algorithm using a convolutional neural network (CNN) to identify low H I column density Lyα absorption systems (log NHI/cm−2 < 17) in the Lyα forest, and predict their physical properties, such as their H I column density (log NHI/cm−2), redshift (zHI), and Doppler width (bHI). Our CNN models are trained using simulated spectra (S/N ≃ 10), and we test their performance on high quality spectra of quasars at redshift z ∼ 2.5 − 2.9 observed with the High Resolution Echelle Spectrometer on the Keck I telescope. We find that ∼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 ≳ 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 ≤ log NHI/cm−2 < 15.5 with a mean absolute error of Δ(log NHI/cm−2) = 0.13, Δ(zHI) = 2.7 × 10−5, and Δ(bHI) = 4.1 km 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 Lyα forest spectra, and thereby greatly increase the statistics of Lyα absorbers.

Citation

Cheng, T.-Y., Cooke, R. J., & Rudie, G. (2022). Harvesting the Lyα forest with convolutional neural networks. Monthly Notices of the Royal Astronomical Society, 517, 755-775. https://doi.org/10.1093/mnras/stac2631

• Journal Article Type: Article
• Acceptance Date: Sep 9, 2022
• Online Publication Date: Sep 17, 2022
• Publication Date: 2022
• Deposit Date: Sep 30, 2022
• Publicly Available Date: Oct 5, 2022
• Journal: Monthly Notices of the Royal Astronomical Society
• Print ISSN: 0035-8711
• Electronic ISSN: 1365-2966
• Publisher: Royal Astronomical Society
• Peer Reviewed: Peer Reviewed
• Volume: 517
• Pages: 755-775
• DOI: https://doi.org/10.1093/mnras/stac2631
• Public URL: https://durham-repository.worktribe.com/output/1190511
• Related Public URLs: https://arxiv.org/abs/2209.02142