🪄 Sandbox

Notebooks, scripts, and tutorials for the SkyExplain project

---

📖 Overview

Sandbox is the tutorial and exploration repository for the SkyExplain project. It hosts Jupyter notebooks and scripts that demonstrate the full analysis pipeline — from CMB map simulation to neural network model selection and interpretability — developed for the paper Explaining Neural Networks on the Sky: Machine Learning Interpretability for CMB Maps.

This repository is intended as a companion to the three core packages, providing worked examples and reproducible figures.

---

📂 Contents

• 🏗️ Architecture notebooks — Explore and test the hybrid PCA–MLP neural network architectures used for CMB model selection, including training configurations and performance diagnostics
• 🔍 SHAP notebooks — Apply SHAP (SHapley Additive exPlanations) to trained network predictions, compute attribution maps, and reproject them onto the HEALPix sphere
• 📊 Plot notebooks — Reproduce some of the figures from the paper, including CMB map visualisations, architecture diagnostics

---

🔗 Related Packages

This repository is part of the SkyExplain ecosystem:

Package	Description
SkySimulation	CMB map simulation (T, Q, U) with ΛCDM and oscillatory features
SkyNeuralNets	NN-based model selection on CMB maps
SkyInterpret	Interpretability analysis of the trained neural networks
Sandbox	Notebooks, scripts, and tutorials (this repo)

---

📦 Installation

Clone the repository and install the required packages:

git clone https://github.com/SkyExplain/Sandbox.git
cd Sandbox
pip install -r requirements.txt

For the full pipeline, also install the core packages:

pip install -e git+https://github.com/SkyExplain/SkySimulation.git
pip install -e git+https://github.com/SkyExplain/SkyNeuralNets.git
pip install -e git+https://github.com/SkyExplain/SkyInterpret.git

---

📄 Citation

If you use this code in your research, please cite:

@article{ocampo2026explaining,
  author        = {Indira Ocampo and Guadalupe Cañas-Herrera},
  title         = {Explaining Neural Networks on the Sky: Machine Learning Interpretability for CMB Maps},
  year          = {2026},
  eprint        = {2604.05290},
  archivePrefix = {arXiv}
}
}

---

📬 Contact

For questions or issues, please open a GitHub Issue or contact indira.ocampo@csic.es.