2026 edition
The 2026 course is in preparation. Registration details, dates, and rooms will be announced here and in the course repository.
The teaching materials below are carried over from 2025 and will be updated as the course is prepared. The previous edition is available in the 2025 course archive.
Marc Lelarge and Tony Bonnaire with Julien Moreau
This course introduces the foundations of machine learning, from statistical models to modern deep learning, with a focus on practical applications in scientific research. Students will learn core methods, computational tools, and workflows to apply machine learning techniques to data and problems in their own field of study.
After completing the core curriculum, each department will supervise (over a six-week period) the projects it has proposed. The purpose of the core curriculum is to provide a solid foundation in the fundamentals of statistical learning, along with the essential computing skills (sklearn – PyTorch) required across all projects.
Sessions
- Session 1: Course Overview - The K-means Algorithm (going further)
- Session 2: Hypothesis Testing
- Session 3: Linear Models on Feature Vectors
- Session 4: Optimization
Practicals
- Practicals 1 - K-Means and SVD
- Practicals 2 - Supervised Learning
- Practicals 3 - Bayes’ Theorem - Naive Bayes Binary Classifier - Logistic Regression from Scratch
- Practicals 4 - PyTorch Tensors 101 - Autograd and Linear Regression