Dataset for: Bayesian deep learning for cosmic volumes with modified gravity

Resumo

The dataset encompasses data utilized in the scientific manuscript titled "Bayesian Deep Learning for Cosmic Volumes with Modified Gravity" available at https://arxiv.org/abs/2309.00612. It includes 3D overdensity fields and power spectra for 2500 simulations of modified gravity (MG) conducted using MG-PICOLA, encompassing 256 Mpc/h side cubical volumes with 128^3 particles. These simulations were employed to derive cosmological parameters through deep neural networks equipped with uncertainty estimations. We evaluated Bayesian neural networks (BNNs) with an enriched approximate posterior distribution, considering two cases: one with a single Bayesian last layer (BLL) and another with Bayesian layers at all levels (FullB). This research contributes to establishing a framework for extracting cosmological parameters from complete, small cosmic volumes extending into the highly nonlinear regime.