Synthetic data experiments for paper on coVariance Neural Networks (NeurIPS 2022).
Jupyter notebook Synthetic_Data_Experiments.ipynb stores the implementation of stability analysis for VNNs and PCA-regression models for synthetic data experiments in Appendix-D of the paper coVariance Neural Networks. The VNNs are based on slight modifications of the GNN libraries from https://github.com/alelab-upenn/graph-neural-networks and these libraries have been included in the Modules and Utils folders as such.
Dependencies:
Python 3.8.5
torch==1.10.2
scikit-learn==0.23.2
pandas==1.1.3
numpy==1.19.2
matplotlib==3.3.2