This repository contains the results of practical work carried out as part of a course on the use of Probabilistic Models in machine learning.
- The first work is an introduction to Bayesian classification, with an application to signal processing
- The second work focuses on image denoising by modelling the original and noisy images as independent or as observations of a Hidden Markov Model.
- The third work is a more open study of an Energy-Based Model, the Restricted Boltzmann Machine, of which the generative properties are studied on the MNIST dataset.