Cosmic topology. Part IVa. Classification of manifolds using machine learning: a case study with small toroidal universes
Co-author Published
JCAP 09 (2024) 057 · 2024
- In brief
- Tests whether machine-learning classifiers can tell small toroidal universes apart from CMB maps, in a setting where the optimal answer can also be computed.
- Details
- Classifiers trained on simulated CMB maps approach likelihood-level discrimination at a fraction of the cost; because the likelihood is computable here, the learned model's gap to the optimal decision rule can be measured.
- My contribution
- Contributed to benchmarking learned topology classification against a directly computable likelihood, and helped prepare and revise the manuscript.
BibTeX
% Abbreviated author list; see the DOI or INSPIRE-HEP for the full record.
@article{Tamosiunas:2024iva,
author = {Tamosiunas, Andrius and Cornet-Gomez, Fernando and Akrami, Yashar and Anselmi, Stefano and {Carr{\'o}n Duque}, Javier and Copi, Craig J. and Eskilt, Johannes R. and G{\"u}ng{\"o}r, {\"O}zen{\c{c}} and Jaffe, Andrew H. and Kosowsky, Arthur and others},
title = {Cosmic topology. {Part IVa}. {Classification} of manifolds using machine learning: a case study with small toroidal universes},
journal = {Journal of Cosmology and Astroparticle Physics},
volume = {2024},
number = {09},
pages = {057},
year = {2024},
doi = {10.1088/1475-7516/2024/09/057},
eprint = {2404.01236},
archivePrefix = {arXiv},
primaryClass = {astro-ph.CO}
}