Separating the EoR Signal with a Convolutional Denoising Autoencoder: A Deep-learning-based Method

Abstract

When applying the foreground removal methods to uncover the faint cosmological signal from the epoch of reionization (EoR), the foreground spectra are assumed to be smooth. However, this assumption can be seriously violated in practice since the unresolved or mis-subtracted foreground sources, which are further complicated by the frequency-dependent beam effects of interferometers, will generate significant fluctuations along the frequency dimension. To address this issue, we propose a novel deep-learning-based method that uses a 9-layer convolutional denoising autoencoder (CDAE) to separate the EoR signal. After being trained on the SKA images simulated with realistic beam effects, the CDAE achieves excellent performance as the mean correlation coefficient ($\bar{\rho }$) between the reconstructed and input EoR signals reaches $0.929\pm 0.045$. In comparison, the two representative traditional methods, namely the polynomial fitting method and the continuous wavelet transform method, both have difficulties in modelling and removing the foreground emission complicated with the beam effects, yielding only ${\bar{\rho }}_{\text{poly}}=0.296\pm 0.121$ and ${\bar{\rho }}_{\text{cwt}}=0.198\pm 0.160$, respectively. We conclude that, by hierarchically learning sophisticated features through multiple convolutional layers, the CDAE is a powerful tool that can be used to overcome the complicated beam effects and accurately separate the EoR signal. Our results also exhibit the great potential of deep-learning-based methods in future EoR experiments.

10 pages, 9 figures; minor text updates to match the MNRAS published version

References in corpus (38)

• Adam: A Method for Stochastic Optimization
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Striving for Simplicity: The All Convolutional Net
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Going Deeper with Convolutions
• WSClean: an implementation of a fast, generic wide-field imager for radio astronomy
• A Full-Sky H-alpha Template for Microwave Foreground Prediction
• Reionization and Cosmology with 21 cm Fluctuations
• Ionospheric Calibration of Low Frequency Radio Interferometric Observations using the Peeling Scheme: I. Method Description and First Results
• Bright Source Subtraction Requirements For Redshifted 21 cm Measurements
• Reionization and the Cosmic Dawn with the Square Kilometre Array
• A semi-empirical simulation of the extragalactic radio continuum sky for next generation radio telescopes
• A Per-Baseline, Delay-Spectrum Technique for Accessing the 21cm Cosmic Reionization Signature
• Foreground simulations for the LOFAR - Epoch of Reionization Experiment
• Towards a free-free template for CMB foregrounds
• Real-Time Calibration of the Murchison Widefield Array
• Foreground Removal using FastICA: A Showcase of LOFAR-EoR
• Constraining the unexplored period between reionization and the dark ages with observations of the global 21 cm signal
• First Season MWA EoR Power Spectrum Results at Redshift 7
• Twenty-one centimeter tomography with foregrounds
• Blind foreground subtraction for intensity mapping experiments
• Towards a model of full-sky Galactic synchrotron intensity and linear polarisation: a re-analysis of the Parkes data
• The Scale of the Problem : Recovering Images of Reionization with GMCA
• An Improved Method for 21cm Foreground Removal
• Statistical 21-cm Signal Separation via Gaussian Process Regression Analysis
• The Evolution of 21-cm Structure (EOS): public, large-scale simulations of Cosmic Dawn and Reionization
• Non-parametric foreground subtraction for 21cm epoch of reionization experiments
• Will point sources spoil 21 cm tomography?
• A Machine Learning Based Morphological Classification of 14,245 Radio AGNs Selected From The Best-Heckman Sample
• DeepSource: Point Source Detection using Deep Learning
• Fast Point Spread Function Modeling with Deep Learning
• Foreground removal for Square Kilometre Array observations of the Epoch of Reionization with the Correlated Component Analysis
• The 21-cm Line as a Probe of Reionization
• Exploiting the Potential of Standard Convolutional Autoencoders for Image Restoration by Evolutionary Search
• Exploring the Cosmic Reionization Epoch in Frequency Space: An Improved Approach to Remove the Foreground in 21 cm Tomography
• Bayesian Inference for Radio Observations
• How to Identify and Separate Bright Galaxy Clusters from the Low-frequency Radio Sky?
• The Application of Continuous Wavelet Transform Based Foreground Subtraction Method in 21 cm Sky Surveys

Cited by in corpus (28)

• Simulation-Based Inference of Reionization Parameters From 3D Tomographic 21 cm Lightcone Images
• deep21: a Deep Learning Method for 21cm Foreground Removal
• Comparing Foreground Removal Techniques for Recovery of the LOFAR-EoR 21cm Power Spectrum
• Recovering the Wedge Modes Lost to 21-cm Foregrounds
• Deep learning approach for identification of HII regions during reionization in 21-cm observations
• Machine Learning for Observational Cosmology
• Comparisons between fast algorithms for the continuous wavelet transform and applications in cosmology: the 1D case
• The bispectrum and 21cm foregrounds during the Epoch of Reionization
• Extracting the 21-cm Power Spectrum and the reionization parameters from mock datasets using Artificial Neural Networks
• Machine-learning recovery of foreground wedge-removed 21-cm light cones for high-$z$ galaxy mapping
• Eliminating polarization leakage effect for neutral hydrogen intensity mapping with deep learning
• The GALAH survey: Characterization of emission-line stars with spectral modelling using autoencoders
• Predicting 21cm-line map from Lyman $\alpha$ emitter distribution with Generative Adversarial Networks
• Analysing the Epoch of Reionization with three-point correlation functions and machine learning techniques
• Noise reduction on single-shot images using an autoencoder
• Probing light relics through cosmic dawn
• A Comparative Analysis To Deal With Missing Spectral Information Caused By RFI In Cosmological HI 21CM Observations
• Pre-selection of the Candidate Fields for Deep Imaging of the Epoch of Reionization with SKA1-Low
• Deep learning approach for identification of HII regions during reionization in 21-cm observations -- III. image recovery
• Field-level Reconstruction from Foreground-Contaminated 21-cm Maps
• Impact of Calibration and Position Errors on Astrophysical Parameters of the HI 21cm Signal
• Synthetic Observations with the Square Kilometre Array (SKA) -- development towards an end-to-end pipeline
• An Alcock-Paczynski Test on Reionization Bubbles for Cosmology
• Periodic Astrometric Signal Recovery through Convolutional Autoencoders
• Constraining the [CII] luminosity function from the power spectrum of line-intensity maps at redshift 3.6
• Deep learning with hybrid frequency differencing and principal component analysis for 21-cm foreground and beam mitigation
• Beyond the Power Spectrum: A New Framework for Non-Stationary Fields with Applications to Light-Cone Effects in Line Intensity Mapping
• Nonlinear reconstruction of 21cm global signal from 21cm power spectrum with artificial neural networks