GNN_Notebooks

A collection of Jupyter notebooks and scripts for learning and experimenting with Graph Neural Networks (GNNs) and graph-related algorithms.
The repository includes both introductory material and more advanced experiments such as graph matching using GNN-based models.

Goal: provide practical examples, experiments, and small pipelines for typical graph-related tasks like node classification, graph-level regression, and graph matching.

---

Repository Structure

• Graph Basics and Algorithms.ipynb
  Introduction to graph theory and basic algorithms using Python and NetworkX. Covers graph representations, traversal, and metrics.

• Node Classification.ipynb
  Example notebook demonstrating node classification using GNNs. Includes data loading, preprocessing, model definition, training, and evaluation.

• Graph Level Regression.ipynb
  Experiments focused on predicting graph-level targets. Explores GNN-based regression pipelines.

• graph_matching/
  Contains notebooks and scripts dedicated to graph matching tasks using Graph Attention Networks (GAT).
  Inside this folder you will find:

  • Implementations and experiments for graph matching pipelines.
  • Code for computing similarity/affinity matrices and visualizing matching results.
  • A dedicated requirements_graph_note.txt file listing the dependencies required for this environment.

• requirements_graph_note.txt
  Top-level dependencies required to reproduce most experiments in the root notebooks.

• LICENSE
  This project is licensed under GPL-3.0.

---

Installation

1. Clone the repository

git clone https://github.com/mgiorgi13/GNN_Notebooks.git
cd GNN_Notebooks

2. Set up the general GNN environment

python -m venv venv
source venv/bin/activate        # macOS/Linux
venv\Scripts\activate.bat       # Windows

pip install --upgrade pip
pip install -r requirements_graph_note.txt

3. Set up Graph Matching environment

cd graph_matching

python -m venv venv_graph_matching
source venv_graph_matching/bin/activate        # macOS/Linux
venv_graph_matching\Scripts\activate.bat       # Windows

pip install --upgrade pip
pip install -r requirements_graph_note.txt