Software implementation of the randomized GNN model presented in the paper Pyramidal Reservoir Graph Neural Networks by Filippo Maria Bianchi, Claudio Gallicchio, and Alessio Micheli.

Downstream tasks

Graph classification

The graph classification experiments can be reproduced by running the script main_class.py.

Dimensionality reduction and visualization

The embeddings generated by the GNN model can be visualized with the script main_lda.py, which uses Linear Discriminant Analysis (LDA) to project the embeddings into a lower-dimensional space.

Model selection

The script cross_validation.py performs a model selection with nested k-fold, similar to the procedure described in the paper A Fair Comparison of Graph Neural Networks for Graph Classification.

Citation

If you are using this material in your research, please consider citing our paper:

@article{bianchi2022pyramidal,
  title={Pyramidal Reservoir Graph Neural Network},
  author={Bianchi, Filippo Maria and Gallicchio, Claudio and Micheli, Alessio},
  journal={Neurocomputing},
  volume={470},
  pages={389--404},
  year={2022},
  publisher={Elsevier}
}