ForSE: a GAN based algorithm for extending CMB foreground models to sub-degree angular scales

Abstract

We present ForSE (Foreground Scale Extender), a novel Python package which aims at overcoming the current limitations in the simulation of diffuse Galactic radiation, in the context of Cosmic Microwave Background experiments (CMB). ForSE exploits the ability of generative adversarial neural networks (GANs) to learn and reproduce complex features present in a set of images, with the goal of simulating realistic and non-Gaussian foreground radiation at sub-degree angular scales. This is of great importance in order to estimate the foreground contamination to lensing reconstruction, de-lensing and primordial B-modes, for future CMB experiments. We applied this algorithm to Galactic thermal dust emission in both total intensity and polarization. Our results show how ForSE is able to generate small scale features (at 12 arc-minutes) having as input the large scale ones (80 arc-minutes). The injected structures have statistical properties, evaluated by means of the Minkowski functionals, in good agreement with those of the real sky and which show the correct amplitude scaling as a function of the angular dimension. The obtained thermal dust Stokes Q and U full sky maps as well as the ForSE package are publicly available for download.

Version 2

References in corpus (7)

• Three Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Polarization Analysis
• A Joint Analysis of BICEP2/Keck Array and Planck Data
• Planck constraints on the tensor-to-scalar ratio
• Spatial variations in the spectral index of polarized synchrotron emission in the 9 yr WMAP sky maps
• B polarization of cosmic microwave background as a tracer of strings
• Cosmic Strings
• Forecasting performance of CMB experiments in the presence of complex foreground contaminations

Cited by in corpus (19)

• A new approach for the statistical denoising of Planck interstellar dust polarization data
• Recovering the CMB Signal with Machine Learning
• Non-Gaussian modelling and statistical denoising of Planck dust polarization full-sky maps using scattering transforms
• MillimeterDL: Deep Learning Simulations of the Microwave Sky
• Full-sky Models of Galactic Microwave Emission and Polarization at Sub-arcminute Scales for the Python Sky Model
• Simulated catalogs and maps of radio galaxies at millimeter wavelengths in Websky
• In-flight polarization angle calibration for LiteBIRD: blind challenge and cosmological implications
• Minkowski Functionals of CMB polarisation intensity with Pynkowski: theory and application to Planck and future data
• Full-sky, arcminute-scale, 3D models of Galactic microwave foreground dust emission based on filaments
• CENN: A fully convolutional neural network for CMB recovery in realistic microwave sky simulations
• Recovering Cosmic Microwave Background Polarization Signals with Machine Learning
• Generative Models of Multi-channel Data from a Single Example -- Application to Dust Emission
• Inference of the optical depth to reionization $\tau$ from $\text{Planck}$ CMB maps with convolutional neural networks
• A convolutional-neural-network estimator of CMB constraints on dark matter energy injection
• Impact of Galactic non-Gaussian foregrounds on CMB lensing measurements
• ForSE+: Simulating non-Gaussian CMB foregrounds at 3 arcminutes in a stochastic way based on a generative adversarial network
• Constraining the polarisation flux density and angle of point sources by training a convolutional neural network
• Simulating a full-sky high resolution Galactic synchrotron spectral index map using neural networks
• Extending Galactic foreground emission with neural networks