Full-sky Cosmic Microwave Background Foreground Cleaning Using Machine Learning

Abstract

In order to extract cosmological information from observations of the millimeter and submillimeter sky, foreground components must first be removed to produce an estimate of the cosmic microwave background (CMB). We developed a machine-learning approach for doing so for full-sky temperature maps of the millimeter and submillimeter sky. We constructed a Bayesian spherical convolutional neural network architecture to produce a model that captures both spectral and morphological aspects of the foregrounds. Additionally, the model outputs a per-pixel error estimate that incorporates both statistical and model uncertainties. The model was then trained using simulations that incorporated knowledge of these foreground components that was available at the time of the launch of the Planck satellite. On simulated maps, the CMB is recovered with a mean absolute difference of <4 μK over the full sky after masking map pixels with a predicted standard error of >50 μK; the angular power spectrum is also accurately recovered. Once validated with the simulations, this model was applied to Planck temperature observations from its 70 GHz through 857 GHz channels to produce a foreground-cleaned CMB map at a Healpix map resolution of nside = 512. Furthermore, we demonstrate the utility of the technique for evaluating how well different simulations match observations, particularly in regard to the modeling of thermal dust.

Machine Learning and Cosmology

“…Although the primary Cosmic Microwave Background (CMB) signal in the standard cosmological scenario can be statistically described as a Gaussian random field and efficiently analyzed with angular power spectrum estimators, machine learning methods have demonstrated superior performance on CMB analysis tasks including CMB lensing reconstruction [41], foreground separation [42,43] and inference with polarization maps [44,45]. Given the anticipated complexity of data, these techniques can significantly enhance the deliverable output of forthcoming surveys like Simons Observatory [46], CMB-S4 [47], and proposed high-resolution experiments like CMB-HD [48].…”

Observational Cosmology with Artificial Neural Networks

“…In addition, ANNs have decreased the computational time taken for cosmological calculations [17][18][19][20][21]. Furthermore, neural networks have made it possible to analyse CMB signals [22,23], and to classify observational measurements from extensive surveys-for example, quasars in the Sloan Digital Sky Survey (SDSS) [24]. Finally, the use of deep learning in cosmology has increased considerably in recent years, and several works have already incorporated this type of research, i.e., [25][26][27][28].…”

Estimation of Full Sky Power Spectrum between Intermediate to Large Angular Scales from Partial Sky CMB Anisotropies using Artificial Neural Network

“…Baccigalupi et al (2000) separated different types of foreground signal, such as thermal dust emissions, galactic synchrotron and radiation emitted by galaxy clusters from CMB maps using ANN. Petroff et al (2020) implemented a neural network to classify the noises from the anisotropies of CMB temperatures.…”