Saoulis, AA;
Piras, D;
Jeffrey, N;
Spurio Mancini, A;
Ferreira, AMG;
Joachimi, B;
(2025)
Transfer learning for multifidelity simulation-based inference in cosmology.
Monthly Notices of the Royal Astronomical Society
, 542
(4)
, Article staf1436. 10.1093/mnras/staf1436.
Preview |
Text
Joachimi_staf1436.pdf Download (2MB) | Preview |
Abstract
Simulation-based inference (SBI) enables cosmological parameter estimation when closed-form likelihoods or models are unavailable. However, SBI relies on machine learning for neural compression and density estimation. This requires large training data sets which are prohibitively expensive for high-quality simulations. We overcome this limitation with multifidelity transfer learning, combining less expensive, lower fidelity simulations with a limited number of high-fidelity simulations. We demonstrate our methodology on dark matter density maps from two separate simulation suites in the hydrodynamical CAMELS Multifield Data set. Pre-training on dark-matter-only N-body simulations reduces the required number of high-fidelity hydrodynamical simulations by a factor between 8 and 15, depending on the model complexity, posterior dimensionality, and performance metrics used. By leveraging cheaper simulations, our approach enables performant and accurate inference on high-fidelity models while substantially reducing computational costs.
| Type: | Article |
|---|---|
| Title: | Transfer learning for multifidelity simulation-based inference in cosmology |
| Open access status: | An open access version is available from UCL Discovery |
| DOI: | 10.1093/mnras/staf1436 |
| Publisher version: | https://doi.org/10.1093/mnras/staf1436 |
| Language: | English |
| Additional information: | © The Author(s) 2025. Published by Oxford University Press on behalf of Royal Astronomical Society. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. |
| Keywords: | Methods: statistical, software: machine learning, cosmology: cosmological parameters, dark matter, large-scale structure of Universe |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > Dept of Physics and Astronomy |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10217022 |
Archive Staff Only
 |
View Item |
