DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks
Abstract
Nextgeneration 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 tensortoscalar ratio r. Achieving competitive constraints on these parameters requires high signaltonoise 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 signaltonoise 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 formore »
 Authors:

 Chicago U., EFI
 Chicago U., KICP
 Chicago U., Astron. Astrophys. Ctr.
 Brown U. (main)
 Descartes Labs, Santa Fe
 Publication Date:
 Research Org.:
 Fermi National Accelerator Lab. (FNAL), Batavia, IL (United States)
 Sponsoring Org.:
 USDOE Office of Science (SC), High Energy Physics (HEP) (SC25)
 OSTI Identifier:
 1487051
 Report Number(s):
 arXiv:1810.01483; FERMILABPUB18515ACD
oai:inspirehep.net:1696852
 Grant/Contract Number:
 AC0207CH11359
 Resource Type:
 Journal Article: Accepted Manuscript
 Journal Name:
 Astron.Comput.
 Additional Journal Information:
 Journal Volume: 28
 Country of Publication:
 United States
 Language:
 English
 Subject:
 79 ASTRONOMY AND ASTROPHYSICS
Citation Formats
Caldeira, J., Wu, W. L.K., Nord, B., Avestruz, C., Trivedi, S., and Story, K. T. DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks. United States: N. p., 2019.
Web. doi:10.1016/j.ascom.2019.100307.
Caldeira, J., Wu, W. L.K., Nord, B., Avestruz, C., Trivedi, S., & Story, K. T. DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks. United States. doi:10.1016/j.ascom.2019.100307.
Caldeira, J., Wu, W. L.K., Nord, B., Avestruz, C., Trivedi, S., and Story, K. T. Wed .
"DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks". United States. doi:10.1016/j.ascom.2019.100307.
@article{osti_1487051,
title = {DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks},
author = {Caldeira, J. and Wu, W. L.K. and Nord, B. and Avestruz, C. and Trivedi, S. and Story, K. T.},
abstractNote = {Nextgeneration 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 tensortoscalar ratio r. Achieving competitive constraints on these parameters requires high signaltonoise 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 signaltonoise 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 continuousvariable 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 hardtomodel nonGaussian astrophysical foreground and noise contributions.},
doi = {10.1016/j.ascom.2019.100307},
journal = {Astron.Comput.},
number = ,
volume = 28,
place = {United States},
year = {2019},
month = {7}
}