nnero 1.0.2

Neural-Network emulator for Reionization and Optical depth

NNERO

This is NNERO (Neural Network Emulator for Reionization and Optical depth), a fast adaptative tool to emulate reionization history using a simple neural network architecture.

The current default networks implemented have been trained on data generated with 21cmCLAST.

---

This package is part of a set of codes which can be combined together to produce forecast or constraints from late-time Universe observables (such as 21cm) on exotic scearios of dark matter and more. Some of these packages are forks of previously existing repositories, some have been written from scratch

• 21cmCLAST forked from 21cmFAST
• HYREC-2 forked from this repository
• MontePython forked from this repository
• 21cmCAST

How to install NNERO?

NNERO can be installed using pip with the following command

pip install nnero

For a manual installation or development you can clone this repository and install it with

git clone https://github.com/gaetanfacchinetti/NNERO.git 
pip install -e .

How to use NNERO?

• A detailed documentation is under construction here.
• A short tutorial can either be found in the documentation or on the wiki page.

Contributions

Any comment or contribution to this project is welcome.

Credits

If you use NNERO or the default classifiers / regressor trained using 21cmCLAST please cite at least one of the following paper that is relevant to your usage:

• G. Facchinetti, Teaching reionization history to machines: \ constraining new physics with early- and late-time probes (in prep.)
• V. Dandoy, C. Doering, G. Facchinetti, L. Lopez-Honorez, J. R. Schwagereit (in prep.)
• G. Facchinetti, A. Korochkin, L. Lopez-Honorez (in prep.)