Unlocking 21cm Cosmology with SBI: A Beginner friendly NRE for Inference of Astrophysical Parameters

Abstract

The 21-cm line of neutral hydrogen is a promising probe of the early Universe, yet extracting astrophysical parameters from its power spectrum remains a major challenge. We present a beginner-friendly PyTorch pipeline for Marginal Neural Ratio Estimation (MNRE), a Simulation-Based Inference (SBI) method that bypasses explicit likelihoods. Using 21cmFAST simulations, we show that MNRE can recover key astrophysical parameters such as the ionizing efficiency $\zeta$ and X-ray luminosity $L_X$ directly from power spectra. Our implementation prioritizes transparency and accessibility, offering a practical entry point for new researchers in 21-cm cosmology.

arXiv commentThis paper has a critical error in the definition of SBI. Even it states explicit use of SBI the paper is rudimentary and inconclusive. This paper was done as a part of undergraduate project so is immature