# Transfer learning for multifidelity simulation-based inference in cosmology

arXiv: https://arxiv.org/abs/2505.21215 · HTML: https://mlsanity.com/paper/2505.21215/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 0 of its 10 references that are in the map.
- First posted on arXiv 27 May 2025.

## Built on by (most cited first, 4 of 4)

- [Multilevel neural simulation-based inference](https://mlsanity.com/paper/2506.06087/) (2025)
- [Simulation-Efficient Cosmological Inference with Multi-Fidelity SBI](https://mlsanity.com/paper/2507.00514/) (2025)
- [Transfer Learning Beyond the Standard Model](https://mlsanity.com/paper/2510.19168/) (2025)
- [Conservative neural posterior estimation via distributionally robust training](https://mlsanity.com/paper/2605.28516/) (2026)

## Abstract (arXiv)

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.