From starlight to dark matter: a stochastic interpolation approach to map dark matter from stellar density
Abstract
The dark matter halo profile in galaxies holds key information about the nature of dark matter and galaxy formation. Constraining the dark matter profile of galaxies beyond the Milky Way traditionally requires expensive spectroscopic observations for kinematic information. In this paper, we explore a conditional generative model framework to map the dark matter profile of Milky Way-mass galaxies from stellar density profiles, more easily obtainable through large photometric imaging surveys. As a proof of concept, we train the model to learn a stochastic bridge between instrument systematics-free baryonic stellar distributions and underlying dark matter density maps from the DREAMS hydrodynamics simulation suite. We recover 2D dark matter density profiles with a typical accuracy of ∼0.1 dex (∼1.5% of the truth in log scale). The stochastic sampling procedure provides uncertainty estimates of the predicted dark matter map, with typical values ∼0.1 dex. Out-of-domain tests with Milky Way-mass galaxies from IllustrisTNG and FIRE simulations show that, while the model can qualitatively be generalized to TNG50 galaxies from IllustrisTNG, the model is sensitive to the galaxy formation model, with ∼0.2-$0.4$ dex over-prediction for the inner profiles (r≲5 kpc) of the FIRE test galaxies. Future work will explore training with additional suites of simulations and/or conditioning on additional information, such as multi-band images. Our results are a first step towards using generative models as a flexible, uncertainty-aware framework for turning forthcoming data from large imaging surveys into spatially resolved dark matter maps.
Related
- Enhanced density fluctuations: consequences for stellar dynamics and dark matter substructure
- Impact of LSST systematics on stellar-stream density fluctuations for dark matter
- Constraints on Dark Matter Annihilation from Starlight-Induced Inverse Compton Emission in Omega Centauri
- Constraining the shape of dark matter haloes using only starlight
- The stability of radially anisotropic rotating stellar systems with a central density cusp
BibTeX
@misc{ou2026from,
title = {{From starlight to dark matter: a stochastic interpolation approach to map dark matter from stellar density}},
author = {Xiaowei Ou and Lina Necib and Carolina Cuesta-Lazaro and Paul Torrey and Niusha Ahvazi and Alyson M. Brooks and Berthy T. Feng and Alex M. Garcia and Jiaxuan Li and Jonah C. Rose and Xuejian Shen and Mark Vogelsberger},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.27028},
eprint = {2609.27028},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.27028},
}
RIS
TY - UNPB
TI - From starlight to dark matter: a stochastic interpolation approach to map dark matter from stellar density
AU - Xiaowei Ou
AU - Lina Necib
AU - Carolina Cuesta-Lazaro
AU - Paul Torrey
AU - Niusha Ahvazi
AU - Alyson M. Brooks
AU - Berthy T. Feng
AU - Alex M. Garcia
AU - Jiaxuan Li
AU - Jonah C. Rose
AU - Xuejian Shen
AU - Mark Vogelsberger
PY - 2026
DA - 2026/09/22
PB - arXiv
DO - 10.48550/arxiv.2609.27028
UR - https://arxiv.org/abs/2609.27028
AB - The dark matter halo profile in galaxies holds key information about the nature of dark matter and galaxy formation. Constraining the dark matter profile of galaxies beyond the Milky Way traditionally requires expensive spectroscopic observations for kinematic information. In this paper, we explore a conditional generative model framework to map the dark matter profile of Milky Way-mass galaxies from stellar density profiles, more easily obtainable through large photometric imaging surveys. As a proof of concept, we train the model to learn a stochastic bridge between instrument systematics-free baryonic stellar distributions and underlying dark matter density maps from the DREAMS hydrodynamics simulation suite. We recover 2D dark matter density profiles with a typical accuracy of ∼0.1 dex (∼1.5% of the truth in log scale). The stochastic sampling procedure provides uncertainty estimates of the predicted dark matter map, with typical values ∼0.1 dex. Out-of-domain tests with Milky Way-mass galaxies from IllustrisTNG and FIRE simulations show that, while the model can qualitatively be generalized to TNG50 galaxies from IllustrisTNG, the model is sensitive to the galaxy formation model, with ∼0.2-$0.4$ dex over-prediction for the inner profiles (r≲5 kpc) of the FIRE test galaxies. Future work will explore training with additional suites of simulations and/or conditioning on additional information, such as multi-band images. Our results are a first step towards using generative models as a flexible, uncertainty-aware framework for turning forthcoming data from large imaging surveys into spatially resolved dark matter maps.
ER -