COSMOGLOBE: Towards end-to-end CMB cosmological parameter estimation without likelihood approximations
Year: 2023
Type: article
Abstract: We implement support for a cosmological parameter estimation algorithm in Commander and quantify its computational efficiency and cost. For a semi-realistic simulation similar to Planck LFI 70 GHz, we find that the computational cost of producing one single sample is about 20 CPU-hours and that the typical Markov chain correlation length is ∼100 samples. The net effective cost per independent sample is ∼2000 CPU-hours, in comparison with all low-level processing costs of 812 CPU-hours for Planck LFI and WMAP in COSMOGLOBE Data Release 1. Thus, although technically possible to run already in its current state, future work should aim to reduce the effective cost per independent sample by one order of magnitude to avoid excessive runtimes, for instance through multi-grid preconditioners and/or derivative-based Markov chain sampling schemes. This work demonstrates the computational feasibility of true Bayesian cosmological parameter estimation with end-to-end error propagation for high-precision CMB experiments without likelihood approximations, but it also highlights the need for additional optimizations before it is ready for full production-level analysis.
Source: Astronomy and Astrophysics
Authors: J. R. Eskilt, K. Lee, D. J. Watts, V. Anshul, R. Aurlien, A. Basyrov, M. Bersanelli, L. P. L. Colombo, H. K. Eriksen, K. S. F. Fornazier, U. Fuskeland, M. Galloway, E. Gjerløw, L. T. Hergt, H. T. Ihle, J. G. S. Lunde, A. Marins, S. K. Nerval, S. Paradiso, F. Rahman, M. San, N.-O. Stutzer, I. K. Wehus
Institutions: University of Helsinki, University of Oslo, University of Milan, Imperial College London, Istituto Nazionale di Fisica Nucleare, Sezione di Milano, Indian Institute of Technology BHU, National Institute for Astrophysics, Universidade de São Paulo, Trieste Astronomical Observatory, University of British Columbia, Planetek Hellas (Greece), California Institute of Technology, Jet Propulsion Laboratory, University of Toronto, Princeton University, University of Waterloo, Helsinki Institute of Physics, Indian Institute of Astrophysics, Lawrence Berkeley National Laboratory
Language: English
Subfield: Astronomy and Astrophysics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,623
Funders: U.S. Department of Energy, National Aeronautics and Space Administration, California Institute of Technology, Ministerio de Ciencia, Innovación y Universidades, National Energy Research Scientific Computing Center, European Space Agency, Government of Canada, European Commission, Deutsche Forschungsgemeinschaft, Ministry of Education, Culture, Sports, Science and Technology, Partnership for Advanced Computing in Europe AISBL, Tekes, Magnus Ehrnroothin Säätiö, China Scholarship Council, Office of Science, Jet Propulsion Laboratory, Goddard Space Flight Center, Natural Sciences and Engineering Research Council of Canada, Japan Society for the Promotion of Science, Agencia Estatal de Investigación, Astrophysics Science Division
Awards: G122673668, G1286236842, G1632040200, G1677143136, G1897041155, 産業モデルの国際的リンケージに関する計量経済学的研究, G2676722403, G3083819904, G3648211085, G3685081974, G3991318801, G4061288673, G4066435496, CMB観測における前景放射の軽減によるインフレーションの発見感度の向上, G419881185, G42916606, G434337789, Beyond Planck -- delivering state-of-the-art observations of the microwave sky from 30 to 70 GHz for the next decade, EXC 2094: ORIGINS: Vom Ursprung des Universums bis zu den ersten Bausteinen des Lebens, G4501827968, G498139845, G5076365615, G5217172575, G526230902, G5555131222, G6337622905, G6480333962, G6902702339, G7866295892, G7942188893, Advanced Methodologies for Next Generation Large Scale CMB Polarization Analysis, G8156463238, G830101876, G8586532580, G8670446301, G8684520763, G969889393
OpenAlex ID: W4386441427