Extracting the 21-cm Power Spectrum and the reionization parameters from mock datasets using Artificial Neural Networks

Detection of the HI 21-cm power spectrum is one of the key science drivers of several ongoing and upcoming low-frequency radio interferometers. However, the major challenge in such observations come from bright foregrounds, whose accurate removal or avoidance is key to the success of these experiments. In this work, we demonstrate the use of artificial neural networks (ANNs) to extract the HI 21-cm power spectrum from synthetic datasets and extract the reionization parameters from the HI 21-cm power spectrum. For the first time, using a suite of simulations, we present an ANN based framework capable of extracting the HI signal power spectrum directly from the total observed sky power spectrum (which contains the 21-cm signal, along with the foregrounds and effects of the instrument). We have used a combination of two ANNs sequentially. In the first step, ANN1 predicts the 21-cm power spectrum directly from foreground corrupted synthetic datasets. In the second step, ANN2 predicts the reionization parameters from the predicted Hi power spectra from ANN1. The two-step ANN framework can be used as an alternative method to extract the 21-cm power spectrum and the reionization parameters directly from foreground dominated datasets. Our ANN-based framework is trained at a redshift of 9 :01, and for k-modes in the range, 0 :17 < k < 0 : 37 Mpc(-1). We have tested the network's performance with mock datasets corrupted with thermal noise corresponding to 1080 hrs of observations of the SKA-1 LOW and HERA. We have recovered the HI power spectra from foreground dominated synthetic datasets, with an accuracy of approximate to 95 99%. We have achieved an accuracy of approximate to 81 - 90% and approximate to 50 - 60% for the predicted reionization parameters, for test sets corrupted with thermal noise corresponding to the SKA-1 LOW and HERA, respectively.