Constraining the Reionization History using Bayesian Normalizing Flows. Issue 3 (21st August 2020)

• Record Type: Journal Article
• Authors: Hortúa, Héctor J.; Malagò, Luigi; Volpi, Riccardo
• Is Part Of: Machine learning: science and technology. Volume 1:Issue 3(2020)
• Publication Date: 2020-08-21
• Subjects: Deep Learning -- Reionization -- Intergalactic Medium -- Cosmology; 006.31
• DOI: 10.1088/2632-2153/aba6f1
• Languages: English
• ISSNs: 2632-2153

Abstract

Abstract: Upcoming experiments such as Hydrogen Epoch of Reionization Array(HERA) and the Square Kilometre Array (SKA) are intended to measure the 21 cm signal over a wide range of redshifts, representing an incredible opportunity in advancing our understanding about the nature of cosmic reionization. At the same time these kind of experiments will present new challenges in processing the extensive amount of data generated, calling for the development of automated methods capable of precisely estimating physical parameters and their uncertainties. In this deliverable we employ Variational Inference, and in particular Bayesian Neural Networks, as an alternative to MCMC in 21 cm observations to report credible estimations for cosmological and astrophysical parameters and assess the correlations among them. Finally, we have implemented the use of bijectors to improve the diagonal Gaussian approximate posteriors and be able to extract significant information from Non-Gaussian signal in the 21 cm dataset.