The Deep Learning Halo Definer: A Multimodal Framework for Halo Mass and Gas Fraction Inference on Galaxy Groups and Clusters
Abstract
Accurately inferring dark matter halo properties like the total halo mass (M_{halo}) and gas fractions (f_{gas}) remains particularly challenging at group scales, where low member counts, shallow potential wells, and AGN feedback-driven baryon expulsion introduce significant observational scatter. As large-scale surveys begin to provide unprecedented multi-wavelength data, there is a pressing need for methods that can jointly leverage diverse observables to overcome these uncertainties. We introduce the Deep Learning Halo Definer (DLHD), a multimodal deep learning framework that simultaneously processes galaxy catalogues and X-ray imaging through the combination of a Deep Sets and a Convolutional Neural Network (CNN) to improve M_{halo} and f_{gas} estimation for galaxy groups and clusters. Using mock datasets derived from the IllustrisTNG300 hydrodynamic simulation, we demonstrate that the DLHD outperforms each of its component networks individually, achieving RMSE improvements in M_{halo} of 1.9x over Deep Sets and 1.3x over the CNN alone. For the gas fraction enclosed within R_{200c}, DLHD reduces RMSE by 2.0x relative to Deep Sets and 1.1x relative to the CNN alone, with consistent improvements across other apertures. These results highlight a novel ability to leverage multi-band information inaccessible to single-modality methods, positioning DLHD as a promising tool for next-generation survey analyses.
Related
- CausticFlow: An Efficient Machine Learning Framework Combining Neural Differential Equations and Normalizing Flows for Binary Microlensing Parameter Inference
- CausticFlow: An Efficient Machine Learning Framework Combining Neural Differential Equations and Normalizing Flows for Binary Microlensing Parameter Inference
- Late-time distance consistency and model-independent curvature inference with strong lensing, quasars, and galaxy clusters
- MUSHY: Multimodal Flow-Based Amortized Bayesian Inference for Spectroscopic Data Fusion
- A Cosmological Framework for Stellar Collisions at High Redshift in Proto-globular Clusters, Nuclear Star Clusters, and Little Red Dots
BibTeX
@misc{ogle2026deep,
title = {{The Deep Learning Halo Definer: A Multimodal Framework for Halo Mass and Gas Fraction Inference on Galaxy Groups and Clusters}},
author = {Caleb Ogle and Kalvyn N. Poncelet Adams and Benjamin D Oppenheimer and Naomi Gluck and Matthew Ho and Daisuke Nagai and Joseph N Burchett and Mohammadreza Ayromlou},
howpublished = {HAL},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.22543},
eprint = {2609.22543},
archivePrefix = {arXiv},
url = {https://hal.science/hal-05785805v1},
}
RIS
TY - UNPB
TI - The Deep Learning Halo Definer: A Multimodal Framework for Halo Mass and Gas Fraction Inference on Galaxy Groups and Clusters
AU - Caleb Ogle
AU - Kalvyn N. Poncelet Adams
AU - Benjamin D Oppenheimer
AU - Naomi Gluck
AU - Matthew Ho
AU - Daisuke Nagai
AU - Joseph N Burchett
AU - Mohammadreza Ayromlou
PY - 2026
DA - 2026/09/18
PB - HAL
DO - 10.48550/arxiv.2609.22543
UR - https://hal.science/hal-05785805v1
AB - Accurately inferring dark matter halo properties like the total halo mass (M_{halo}) and gas fractions (f_{gas}) remains particularly challenging at group scales, where low member counts, shallow potential wells, and AGN feedback-driven baryon expulsion introduce significant observational scatter. As large-scale surveys begin to provide unprecedented multi-wavelength data, there is a pressing need for methods that can jointly leverage diverse observables to overcome these uncertainties. We introduce the Deep Learning Halo Definer (DLHD), a multimodal deep learning framework that simultaneously processes galaxy catalogues and X-ray imaging through the combination of a Deep Sets and a Convolutional Neural Network (CNN) to improve M_{halo} and f_{gas} estimation for galaxy groups and clusters. Using mock datasets derived from the IllustrisTNG300 hydrodynamic simulation, we demonstrate that the DLHD outperforms each of its component networks individually, achieving RMSE improvements in M_{halo} of 1.9x over Deep Sets and 1.3x over the CNN alone. For the gas fraction enclosed within R_{200c}, DLHD reduces RMSE by 2.0x relative to Deep Sets and 1.1x relative to the CNN alone, with consistent improvements across other apertures. These results highlight a novel ability to leverage multi-band information inaccessible to single-modality methods, positioning DLHD as a promising tool for next-generation survey analyses.
ER -