Super-resolving Polarized Dust Emission with Transformer-Based Multi-Tracer Fusion
Abstract
Realistic models of polarized Galactic dust emission at high angular resolution are essential for component separation in next-generation cosmic microwave background (CMB) B-mode experiments. However, existing dust polarization templates are constrained only to ~1 degrees, and it remains challenging to develop models that can generate realistic small-scale complexity that is also coherently connected to the well-measured large-scale structures. In this work, we develop a data-driven super-resolution framework that predicts small-scale dust polarization structure directly from multiple tracers observed at high resolution. We combine the Planck dust optical depth τ353, \ion{H}{1}-based Stokes templates, and coarse Planck GNILC Q and U maps in a transformer-based model trained to recover native-resolution GNILC polarization from a 4× beam-smoothed version of itself, so that the network solves a deconvolution problem informed by the structure of multiple tracers rather than generating structure to match a target statistic. Applying the model to the uniform-resolution GNILC maps, we produce Q and U predictions at 20' over the high-latitude (|b|>30 degrees) sky. Using scalar and tensorial Minkowski functionals and scattering transform statistics, we find that our maps exhibit more coherent, anisotropic, and filamentary small-scale structure than the latest PySM dust model. Our results establish data-driven learned deconvolution as a promising route to foreground models whose small scales are inferred from data rather than prescribed.
Related
- H2O emission as tracer of pebble drift: Insights from coupling transport and thermochemical models
- Vacuum birefringence and the polarized X-ray emission from a radio magnetar
- Far-Ultraviolet diffuse emission as a tracer of Galactic extinction at the North Galactic Pole
- H2O emission as tracer of pebble drift: insights from coupling transport and thermochemical models
- Simulating and Characterizing Non-Gaussian Small-Scale Polarized Emission
BibTeX
@misc{lei2026superresolving,
title = {{Super-resolving Polarized Dust Emission with Transformer-Based Multi-Tracer Fusion}},
author = {Minjie Lei and George Halal and S. E. Clark and Alejandro Dobles and Katie Brown and Viraj Manwadkar},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.02575},
eprint = {2610.02575},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.02575},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Super-resolving Polarized Dust Emission with Transformer-Based Multi-Tracer Fusion
AU - Minjie Lei
AU - George Halal
AU - S. E. Clark
AU - Alejandro Dobles
AU - Katie Brown
AU - Viraj Manwadkar
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.02575
UR - https://arxiv.org/abs/2610.02575
AB - Realistic models of polarized Galactic dust emission at high angular resolution are essential for component separation in next-generation cosmic microwave background (CMB) B-mode experiments. However, existing dust polarization templates are constrained only to ~1 degrees, and it remains challenging to develop models that can generate realistic small-scale complexity that is also coherently connected to the well-measured large-scale structures. In this work, we develop a data-driven super-resolution framework that predicts small-scale dust polarization structure directly from multiple tracers observed at high resolution. We combine the Planck dust optical depth τ_353, \ion{H}{1}-based Stokes templates, and coarse Planck GNILC Q and U maps in a transformer-based model trained to recover native-resolution GNILC polarization from a 4× beam-smoothed version of itself, so that the network solves a deconvolution problem informed by the structure of multiple tracers rather than generating structure to match a target statistic. Applying the model to the uniform-resolution GNILC maps, we produce Q and U predictions at 20' over the high-latitude (|b|>30 degrees) sky. Using scalar and tensorial Minkowski functionals and scattering transform statistics, we find that our maps exhibit more coherent, anisotropic, and filamentary small-scale structure than the latest PySM dust model. Our results establish data-driven learned deconvolution as a promising route to foreground models whose small scales are inferred from data rather than prescribed.
ER -