SRoll3: A neural network approach to reduce large-scale systematic effects in the Planck High Frequency Instrument maps

Abstract

In the present work, we propose a neural network based data inversion approach to reduce structured contamination sources, with a particular focus on the mapmaking for Planck High Frequency Instrument (Planck-HFI) data and the removal of large-scale systematic effects within the produced sky maps. The removal of contamination sources is rendered possible by the structured nature of these sources, which is characterized by local spatiotemporal interactions producing couplings between different spatiotemporal scales. We focus on exploring neural networks as a means of exploiting these couplings to learn optimal low-dimensional representations, optimized with respect to the contamination source removal and mapmaking objectives, to achieve robust and effective data inversion. We develop multiple variants of the proposed approach, and consider the inclusion of physics informed constraints and transfer learning techniques. Additionally, we focus on exploiting data augmentation techniques to integrate expert knowledge into an otherwise unsupervised network training approach. We validate the proposed method on Planck-HFI 545 GHz Far Side Lobe simulation data, considering ideal and non-ideal cases involving partial, gap-filled and inconsistent datasets, and demonstrate the potential of the neural network based dimensionality reduction to accurately model and remove large-scale systematic effects. We also present an application to real Planck-HFI 857 GHz data, which illustrates the relevance of the proposed method to accurately model and capture structured contamination sources, with reported gains of up to one order of magnitude in terms of contamination removal performance. Importantly, the methods developed in this work are to be integrated in a new version of the SRoll algorithm (SRoll3), and we describe here SRoll3 857 GHz detector maps that will be released to the community.

arXiv comment16 pages, 8 figures

References in corpus (22)

• HEALPix -- a Framework for High Resolution Discretization, and Fast Analysis of Data Distributed on the Sphere
• Theory-guided Data Science: A New Paradigm for Scientific Discovery from Data
• Deep learning for universal linear embeddings of nonlinear dynamics
• Hidden Physics Models: Machine Learning of Nonlinear Partial Differential Equations
• Explainable Machine Learning for Scientific Insights and Discoveries
• A Review of Convolutional Neural Networks for Inverse Problems in Imaging
• Adversarial Uncertainty Quantification in Physics-Informed Neural Networks
• Planck 2018 results. III. High Frequency Instrument data processing and frequency maps
• Planck 2015 results. VIII. High Frequency Instrument data processing: Calibration and maps
• Planck 2013 results. VIII. HFI photometric calibration and mapmaking
• Planck 2015 results. III. LFI systematic uncertainties
• A Map-Making Algorithm for the Planck Surveyor
• MAPCUMBA : a fast iterative multi-grid map-making algorithm for CMB experiments
• Intermittent process analysis with scattering moments
• SRoll2: an improved mapmaking approach to reduce large-scale systematic effects in the Planck High Frequency Instrument legacy maps
• ROMA: a map-making algorithm for polarised CMB data sets
• The RWST, a comprehensive statistical description of the non-Gaussian structures in the ISM
• Madam - a map-making method for CMB experiments
• Making CMB temperature and polarization maps with Madam
• Removing 1/f noise stripes in cosmic microwave background anisotropy observations
• Comparison of map-making algorithms for CMB experiments
• PReBeaM for Planck: A Polarized Regularized Beam Deconvolution Map-Making Method

Cited by in corpus (2)

• Non-Gaussian modelling and statistical denoising of Planck dust polarization full-sky maps using scattering transforms
• Performance of the polarization leakage correction in the PILOT data