Signal Processing for Machine Learning — Laboratory Sessions

This repository hosts the laboratory exercises and supporting materials for the Signal Processing for Machine Learning course taught by Prof. Sergio Barbarossa and Prof. Paolo Di Lorenzo , within the Master’s program in Data Science at Sapienza University of Rome.

---

Repository Structure

sparse_learning_lab/

This directory contains the code presented for the first laboratory session dedicated to Linear and deep methods in sparse learning, and solutions to the exercises. You can enjoy the notebook in colab clicking over the badge:

Contents:

• ISTA_LISTA.ipynb — Exercises on sparse learning:

  • Implementation of Iterative Soft Thresholding Algirithm (ISTA);.
  • Implementation of Learnable Iterative Soft Thresholding Algorithm (LISTA);
  • Implementation of Analytical Learnable Iterative Soft Thresholding Algorithm ALISTA).

• ISTA_LISTA_SOLUTIONS.ipynb — Solutions to the proposed exercises.

gnn_lab/

This directory contains the code presented for the second laboratory session dedicated to Graph Neural Networks (GNNs), and solutions to the exercises. You can enjoy the notebook in colab clicking over the badge:

Contents:

• ex1.py — Exercise 1 (in-class)
  Implementation of a simple GCN for semi-supervised node classification.

• ex2.py — Exercise 2 (in-class)
  Implementation of a GCN for graph classification using a global readout.

• home_ex1.py — Home Exercise 1 (solution)
  GCN for graph classification using hierarchical (non-global) pooling layers.

• home_ex2.py — Home Exercise 2 (solution)
  Implementation of a Graph Attention Network (GAT) for node classification.

---