Alejandro Pinto

Bayesian Inference of Primordial Magnetic Field Parameters from CMB with Spherical Graph Neural Networks

J. A. Pinto Castro, H. J. Hortúa, J. E. García-Farieta, R. A. Hurtado

A novel Bayesian graph deep learning framework for estimating key cosmological parameters in a primordial magnetic field (PMF) cosmology from simulated Cosmic Microwave Background (CMB) maps. The methodology uses DeepSphere — a spherical convolutional neural network respecting CMB's spherical geometry via HEALPix pixelization — combined with Bayesian Neural Networks (BNNs) for robust uncertainty quantification. The framework achieves R² scores exceeding 89% for magnetic parameter estimation.