Fondamentaux théoriques du machine learning / Python & machine learning

This repository contains slides and exercises for Epita students of the FTML and PTML courses.

FTML

This folder contains the slides of the courses and some related exercises.

1. Introduction
2. Mathematical tools for ML, Supervised learning,
3. Risk decomposition, Optimization in ML, Ordinary least squares.
4. Ridge regression, Classification, Logistic regression
5. Unsupervised learning: clustering, dimensionality reduction (PCA)
6. Gradient algorithms, probabilistic modelling
7. Support vector machines, Kernel methods
8. Decision tree learning, Ensemble learning
9. Scoring and cross-validation, neural networks I
10. Neural networks II, statistical learning
11. Local methods, Sparse methods and variable selection, adaptivity
12. Bayesian learning, latent variables

PTML

This folder contains the practical sessions.

1. Reminders/Introduction to Python 3, polynomial regression, Ordinary least squares
2. Ridge regression, logistic regression, cross validation I
3. Gradient descent (GD) and extensions (on OLS)
4. Stochastic gradient descent (SGD) and extensions (on logistic regression)
5. Stochastic average gradient (SAG) (on logistic regression)
6. Bayesian learning
7. Model selection, PCA, kernel PCA
8. Neural networks
9. Nonlinear dimensionality reduction
10. Modern topics on gradient descent