DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks

Abstract

Next-generation cosmic microwave background (CMB) experiments will have lower noise and therefore increased sensitivity, enabling improved constraints on fundamental physics parameters such as the sum of neutrino masses and the tensor-to-scalar ratio r. Achieving competitive constraints on these parameters requires high signal-to-noise extraction of the projected gravitational potential from the CMB maps. Standard methods for reconstructing the lensing potential employ the quadratic estimator (QE). However, the QE performs suboptimally at the low noise levels expected in upcoming experiments. Other methods, like maximum likelihood estimators (MLE), are under active development. In this work, we demonstrate reconstruction of the CMB lensing potential with deep convolutional neural networks (CNN) - ie, a ResUNet. The network is trained and tested on simulated data, and otherwise has no physical parametrization related to the physical processes of the CMB and gravitational lensing. We show that, over a wide range of angular scales, ResUNets recover the input gravitational potential with a higher signal-to-noise ratio than the QE method, reaching levels comparable to analytic approximations of MLE methods. We demonstrate that the network outputs quantifiably different lensing maps when given input CMB maps generated with different cosmologies. We also show we can use the reconstructed lensing map for cosmological parameter estimation. This application of CNN provides a few innovations at the intersection of cosmology and machine learning. First, while training and regressing on images, we predict a continuous-variable field rather than discrete classes. Second, we are able to establish uncertainty measures for the network output that are analogous to standard methods. We expect this approach to excel in capturing hard-to-model non-Gaussian astrophysical foreground and noise contributions.

arXiv comment19 pages; LaTeX; 12 figures; changes to match published version

References in corpus (36)

• Planck 2015 results. XIII. Cosmological parameters
• Nine-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Cosmological Parameter Results
• Efficient Computation of CMB anisotropies in closed FRW models
• Road Extraction by Deep Residual U-Net
• Nine-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Final Maps and Results
• The Simons Observatory: Science goals and forecasts
• Weak Gravitational Lensing of the CMB
• Signature of Gravity Waves in Polarization of the Microwave Background
• A high-bias, low-variance introduction to Machine Learning for physicists
• A Probe of Primordial Gravity Waves and Vorticity
• BICEP2 / Keck Array VI: Improved Constraints On Cosmology and Foregrounds When Adding 95 GHz Data From Keck Array
• Mass Reconstruction with CMB Polarization
• BICEP2 / Keck Array x: Constraints on Primordial Gravitational Waves using Planck, WMAP, and New BICEP2/Keck Observations through the 2015 Season
• Weak Lensing of the CMB: A Harmonic Approach
• Mission design of LiteBIRD
• First Estimations of Cosmological Parameters From BOOMERANG
• Reconstruction of lensing from the cosmic microwave background polarization
• Averaging Weights Leads to Wider Optima and Better Generalization
• The Atacama Cosmology Telescope: Two-Season ACTPol Spectra and Parameters
• Measurements of the Temperature and E-Mode Polarization of the CMB from 500 Square Degrees of SPTpol Data
• Measurement of the Cosmic Microwave Background Polarization Lensing Power Spectrum with the POLARBEAR experiment
• DeepSphere: Efficient spherical Convolutional Neural Network with HEALPix sampling for cosmological applications
• Lensing Reconstruction with CMB Temperature and Polarization
• Delensing CMB Polarization with External Datasets
• CNN-based Segmentation of Medical Imaging Data
• A Measurement of the Cosmic Microwave Background Gravitational Lensing Potential from 100 Square Degrees of SPTpol Data
• The Atacama Cosmology Telescope: Two-Season ACTPol Lensing Power Spectrum
• A Measurement of the Cosmic Microwave Background $B$-Mode Polarization Power Spectrum at Sub-Degree Scales from 2 years of POLARBEAR Data
• Measurements of Sub-degree B-mode Polarization in the Cosmic Microwave Background from 100 Square Degrees of SPTpol Data
• A Guide to Designing Future Ground-based CMB Experiments
• BICEP2 / Keck Array VIII: Measurement of gravitational lensing from large-scale B-mode polarization
• A 2500 square-degree CMB lensing map from combined South Pole Telescope and Planck data
• CMB Polarization B-mode Delensing with SPTpol and Herschel
• Bayesian delensing of CMB temperature and polarization
• Clebsch-Gordan Nets: a Fully Fourier Space Spherical Convolutional Neural Network
• Estimating Cosmological Parameters from the Dark Matter Distribution

Cited by in corpus (49)

• The CosmoVerse White Paper: Addressing observational tensions in cosmology with systematics and fundamental physics
• Learning to Predict the Cosmological Structure Formation
• What can Machine Learning tell us about the background expansion of the Universe?
• Bayesian delensing delight: sampling-based inference of the primordial CMB and gravitational lensing
• The Benefits of CMB Delensing
• Parameters Estimation for the Cosmic Microwave Background with Bayesian Neural Networks
• Full-sky Cosmic Microwave Background Foreground Cleaning Using Machine Learning
• A Demonstration of Improved Constraints on Primordial Gravitational Waves with Delensing
• Convolutional Neural Networks on the HEALPix sphere: a pixel-based algorithm and its application to CMB data analysis
• MUSE: Marginal Unbiased Score Expansion and Application to CMB Lensing
• Machine Learning for Observational Cosmology
• Baryon acoustic oscillations reconstruction using convolutional neural networks
• Applications of Machine Learning to Predicting Core-collapse Supernova Explosion Outcomes
• ForSE: a GAN based algorithm for extending CMB foreground models to sub-degree angular scales
• Recovering the CMB Signal with Machine Learning
• Extracting cosmological parameters from N-body simulations using machine learning techniques
• Cleaning our own Dust: Simulating and Separating Galactic Dust Foregrounds with Neural Networks
• MillimeterDL: Deep Learning Simulations of the Microwave Sky
• Cosmic Velocity Field Reconstruction Using AI
• Comparison of delensing methodologies and assessment of the delensing capabilities of future experiments
• LiteBIRD Science Goals and Forecasts: Improving Sensitivity to Inflationary Gravitational Waves with Multitracer Delensing
• Inpainting Galactic Foreground Intensity and Polarization maps using Convolutional Neural Network
• Identification of point sources in gamma rays using U-shaped convolutional neural networks and a data challenge
• Reconstructing Patchy Reionization with Deep Learning
• Foreground model recognition through Neural Networks for CMB B-mode observations
• Connecting optical morphology, environment, and HI mass fraction for low-redshift galaxies using deep learning
• DeepSZ: Identification of Sunyaev-Zel'dovich Galaxy Clusters using Deep Learning
• Reconstructing Cosmic Polarization Rotation with ResUNet-CMB
• Mitigating Foreground Bias to the CMB Lensing Power Spectrum for a CMB-HD Survey
• Inference of gravitational lensing and patchy reionization with future CMB data
• Deep learning the astrometric signature of dark matter substructure
• Deep Forest: Neural Network reconstruction of the Lyman-alpha forest
• Delensing of Cosmic Microwave Background Polarization with machine learning
• Inference of the optical depth to reionization $\tau$ from $\text{Planck}$ CMB maps with convolutional neural networks
• Cluster profiles from beyond-the-QE CMB lensing mass maps
• Lensing reconstruction from the cosmic microwave background polarization with machine learning
• Deep Forest: Neural Network reconstruction of intergalactic medium temperature
• Classifying CMB time-ordered data through deep neural networks
• Exploring fundamental laws of classical mechanics via predicting the orbits of planets based on neural networks
• A convolutional-neural-network estimator of CMB constraints on dark matter energy injection
• Deep Needlet: A CNN based full sky component separation method in Needlet space
• Self-supervised component separation for the extragalactic submillimeter sky
• Datacube segmentation via Deep Spectral Clustering
• Unsupervised Domain Adaptation for Constraining Star Formation Histories
• CMB delensing with deep learning
• Baryon density extraction and isotropy analysis of Cosmic Microwave Background using Deep Learning
• Galaxies OBserved as Low-luminosity Identified Nebulae (GOBLIN): a catalog of 43,000 high-probability dwarf galaxy candidates in the UNIONS survey
• Towards characterizing dark matter subhalo perturbations in stellar streams with graph neural networks
• Power Spectrum Emulators from Neural Networks and Tree-Based Methods