Recovering CMB polarization maps with neural networks: performance in realistic simulations

Recovering the polarized cosmic microwave background (CMB) is essential for shedding light on the exponential expansion of the very early Universe, known as cosmic inflation. Achieving this goal requires not only improved instrumental sensitivity but also the development of robust and diverse data analysis techniques. In this work, we explore a novel component separation approach based on neural networks, previously validated using realisticPlancktemperature simulations, to reconstruct the StokesQandUpolarization maps. To validate the method, we first test the network on realisticPlancksky simulations of regions deliberately excluded from the training set. We compare the input and outputEEandBBpower spectra, finding a mean absolute error of 0.1 ± 0.3μK2for theE-mode and -0.1 ± 0.3μK2for theB-mode. These results demonstrate a partial recovery of theE-mode and a limited recovery of theB-mode, the latter remaining dominated by residualPlancknoise. We then apply the trained network to publicPlanckobservations, recovering CMB polarization maps broadly consistent with those obtained using the Commander method. The recoveredEEspectra differ by less than 5% from the reference at intermediate and small angular scales, although significant discrepancies remain at large scales, which may impact cosmological interpretations. These results, while encouraging, clearly reflect the limitations of the current setup and motivate further improvements in training data and methodology. Based on these findings, we conclude that neural network-based methods show potential as component separation techniques in polarization CMB experiments. However, substantial improvements and more comprehensive analyses are necessary before these methods can provide reliable high-precision cosmological estimates.