leidenuniv.nl, as data

Astrophysics authors/titles "new"The next generation 21 cm surveys open a new window onto the early stages ofcosmic structure formation and provide new insights about the Epoch ofReionization (EoR). However, the non-Gaussian nature of the 21 cm signal alongwith the huge amount of data generated from these surveys will require moreadvanced techniques capable to efficiently extract the necessary information toconstrain the Reionization History of the Universe. In this paper we presentthe use of Bayesian Neural Networks (BNNs) to predict the posteriordistribution for four astrophysical and cosmological parameters. Besidesachieving state-of-the-art prediction performances, the proposed methodsprovide accurate estimation of parameters uncertainties and infer correlationsamong them. Additionally, we demonstrate the advantages of Normalizing Flows(NF) combined with BNNs, being able to model more complex output distributionsand thus capture key information as non-Gaussianities in the parameterconditional density distribution for astrophysical and cosmological dataset.Finally, we propose novel calibration methods employing Normalizing Flows aftertraining, to produce reliable predictions, and we demonstrate the advantages ofthis approach both in terms of computational cost and prediction performances.Title: Effects of Supernova Redshift Uncertainties on the Determination of Cosmological ParametersSubjects: Cosmology and Nongalactic Astrophysics (astro-ph.CO)Redshifts used in current cosmological supernova samples are measured usingtwo primary techniques, one based on well-measured host galaxy spectral linesand the other based on supernova-dominated spectra. Here, we construct anupdated Pantheon catalog with revised redshifts, redshift sources and estimateduncertainties for the entire sample to investigate whether these two techniquesyield consistent results. The best-fit cosmological parameters using these two