Introduction to machine learning

Material

• Lecture notes (September 14 2026)

Practicals

• Functions for practical sessions

• Practical 1

• Practical 2, the SVM GUI and the polynomial kernel in action

• Practical 3

• Practical 4 (supplementary)

• Practical 5 (supplementary)

• Practical 6 (supplementary)

• Practical 7 (supplementary)

Exam (2026-2027)

1. Exam without documents (Thursday, November 5): 16 points.

2. Data challenge: 4 points.

Instructions for the data challenge

1. Join the challenge Maladie de Parkinson : Prédire et Corriger les Biais dans l’Évaluation du Score Moteur

2. Select the course SU - AS 2026

3. Fill out the registration form

4. Intermediate ranking: Sunday, October 18 at 10PM

5. Final ranking and end of the data challenge: Sunday, November 8 at 10PM

Final mark:

1. Intermediate and final rankings on the private leaderboards

2. A notebook making it possible to preprocess the data, train your final model et generate the predictions. To be uploaded on Moodle before the end of the data challenge

3. A 2-page pdf-LaTeX report describing:

• the data preprocessing;

• the model selection phase;

• your final model along with its public and private scores;

• the impediments you faced and your achievements. To be uploaded on the data challenge platform.