BibTeX

@article{2505.21215v2,
Author        = {Alex A. Saoulis and Davide Piras and Niall Jeffrey and Alessio Spurio Mancini and Ana M. G. Ferreira and Benjamin Joachimi},
Title         = {Transfer learning for multifidelity simulation-based inference in
  cosmology},
Eprint        = {2505.21215v2},
DOI           = {10.1093/mnras/staf1436},
ArchivePrefix = {arXiv},
PrimaryClass  = {astro-ph.CO},
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 datasets 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 Dataset. 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.},
Year          = {2025},
Month         = {May},
Note          = {Mon Not R Astron Soc (2025) 3231-3245},
Url           = {http://arxiv.org/abs/2505.21215v2},
File          = {2505.21215v2.pdf}
}