CosmoSim Experiments July 2026

These pages report on three experiments made July 2026.

• Experiment 1 (singleton SIE)

• Experiment 2 (2xSIE)

• Experiment 3 (4xSIS)

All the experiments follow the same general protocol, elaborate in Training and testing a single model., with the following steps.

1. Dataset generation.

2. Training of a single machine learning model.

3. Testing and analysis.

This is based on the general Machine Learning Pipeline.

You can download and run the different documents in Jupyter Lab. Notebooks and datafiles can be downloaded one by one from the pages, or you can clone the entire site from github.

To install the dependencies, download requirements.txt and use

pip install -r requirements.txt

To open the files from git in jupyter lab, they must be converted from md:myst to ipynb by jupytext. If you have make you can use the Makefile provided.

make ipynb

Home to Mapping Dark Matter via Gravitational Lensing