CRANN is an open-source project designed to facilitate cosmological reconstructions using Artificial Neural Networks. This repository hosts the neural networks trained to generate synthetic cosmic chronometers, JLA Type Ia supernovae (distance modulus), and
Gómez-Vargas, I., Medel-Esquivel, R., García-Salcedo, R. et al. Neural network reconstructions for the Hubble parameter, growth rate, and distance modulus. Eur. Phys. J. C 83, 304 (2023). https://doi.org/10.1140/epjc/s10052-023-11435-9.
- arxiv_notebooks/: Contains Jupyter notebooks that reproduce the figures presented in the paper. These notebooks provide insights into the data generation process and the neural network training details.
To use the CRANN models or to contribute to the project, you will need to install several dependencies. Ensure you have the following Python packages installed:
- numpy
- sklearn
- scipy
- pandas
- matplotlib
- seaborn
- tensorflow==2.6.0
- keras==2.6.0
- astroNN
- h5py==2.9.0
You can install these packages using pip
by running:
pip install numpy sklearn scipy pandas matplotlib seaborn tensorflow==2.6.0 keras==2.6.0 astroNN h5py==2.9.0
Details about the training of the models can be found in the arxiv_notebooks/ directory.
If you use CRANN in your research, please cite our paper to acknowledge the work that has gone into developing this resource:
@article{GomezVargas2023,
title={Neural network reconstructions for the Hubble parameter, growth rate and distance modulus},
author={Gómez-Vargas, I. and Medel-Esquivel, R. and García-Salcedo, R. and others},
journal={Eur. Phys. J. C},
volume={83},
pages={304},
year={2023},
publisher={Springer}
}
For more details on the training process and development of these models, please refer to our related repository: neuralCosmoReconstruction.
This repository is currently under construction. We welcome contributions and suggestions to improve the project.