Cosmological simulation suite CosmoGridV1
CosmoGridV1 is a large lightcone simulation set for map-level, simulation-based cosmological inference with probes of large scale structure. It is designed for practical parameter measurement with Stage-III survey data and Machine Learning inference. It contains 2500 unique cosmological models spanning different combinations of 6 cosmological parameters and is publicly available at www.cosmogrid.ai.
Analysing the weak gravitational lensing maps with Artificial Intelligence algorithms can greatly increase the amount of cosmological information recovered from the data. This requires simulated theory prediction for different cosmological models, which we would like to compare with the data using AI. To enable this, we created a large dark matter simulation set, called the CosmoGridV1. It contains 2500 unique cosmological models spanning different combinations of 6 cosmological parameters, which govern the composition of the universe. These simulations can be used to constrain these parameters using large weak lensing surveys. CosmoGrid was ran at the Swiss Supercomputing Center (CSCS) in the large production project “Measuring Dark Energy with Deep Learning”, by a collaboration of researchers at ETH Zurich and the University of Zurich. The data was was described in the paper external page (https://arxiv.org/abs/2209.04662) and made available to the public at the website external page www.cosmogrid.ai.