Extracting the Epoch of Reionization Signal with 3D U-Net Neural Networks Using a Data-driven Systematic Effect Model

Extracting the Epoch of Reionization Signal with 3D U-Net Neural Networks Using a Data-driven Systematic Effect Model

Neutral hydrogen serves as a crucial probe for the Cosmic Dawn and the Epoch of Reionization (EoR). Actual observations of the 21 cm signal often encounter challenges such as thermal noise and various systematic effects. To overcome these challenges, we simulate SKA-Low-depth images in the South Cel...

Full description

Saved in:
Bibliographic Details
Main Authors: Li-Yang Gao, Léon V. E. Koopmans, Florent G. Mertens, Satyapan Munshi, Yichao Li, Stefanie A. Brackenhoff, Emilio Ceccotti, J. Kariuki Chege, Anshuman Acharya, Raghunath Ghara, Sambit K. Giri, Ilian T. Iliev, Garrelt Mellema, Xin Zhang
Format: Article
Language:English
Published: IOP Publishing 2025-01-01
Series:The Astrophysical Journal
Subjects:
H I line emission
Reionization
Gaussian Processes regression
Neural networks
Online Access:https://doi.org/10.3847/1538-4357/ade2dc
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Neutral hydrogen serves as a crucial probe for the Cosmic Dawn and the Epoch of Reionization (EoR). Actual observations of the 21 cm signal often encounter challenges such as thermal noise and various systematic effects. To overcome these challenges, we simulate SKA-Low-depth images in the South Celestial Pole field and process them with a deep learning method. We utilized foreground residuals acquired by LOFAR during actual North Celestial Pole field observations, thermal and excess variances calculated via Gaussian process regression, and 21 cm signals generated with 21cmFAST for signal extraction tests. Our approach to overcome these foreground, thermal noise, and excess variance components employs a 3D U-Net neural network architecture for image analysis. When considering thermal noise corresponding to 1752 hr of integration time, U-Net provides reliable 2D power spectrum predictions, and robustness tests ensure that we get realistic EoR signals. Adding foreground residuals, however, causes inconsistencies below the horizon delay line. Lastly, evaluating both thermal noise and excess variances with observations up to 4380 hr and 13,140 hr ensures reliable power spectrum estimations within the EoR window and across nearly all scales, respectively. The incoherence of excess variances in the frequency direction can greatly affect deep learning to extract 21 cm signals.
ISSN:1538-4357

Similar Items