Sequential Learning (Orsay)

Sequential Learning course for Master 2 Mathématiques de l’Aléatoire & MathIA.

Year 2024-2025

Courses are every Tuesday, 2pm-5pm in room 1A11.

Note: there is no course on October 29 and November 19.

Course Evaluation

• 1 homework
• 1 final written exam.

Each student is allowed a single handwritten, two-sides, cheatsheet for the exam.

The homework is available here and is due for November 13, 6pm.

Lecture Notes

• Lecture 1 (learning with experts)
• Lecture 2 (concentration inequalities)
• Lecture 3 (stochastic bandits, part 1)
• Lecture 4 (stochastic bandits, part 2)
• Lecture 5 (lower bound)

• Lecture 6 (MOSS and continuum of arms)

• Exercise sheet n°1 (lectures 1 to 4)
• Exercise sheet n°2 (lectures 5 to 8)

Previous years material

• Lecture 1 (learning with experts)
• Lecture 2 (concentration inequalities)
• Lecture 3 (stochastic bandits, part 1)
• Lecture 4 (stochastic bandits, part 2)
• Lecture 5 (some properties of the KL)
• Lecture 6 (lower bound)
• Lecture 7 (MOSS and continuum of arms))
• Lecture 8 (contextual, linear bandits)
• Lecture 9 (pure exploration, BAI)

Bibliographic Resources

• Bandit algorithms, Tor Lattimore and Csaba Szepesvári. Cambridge University Press, 2020.
• Introduction to Multi-Armed Bandits, Aleksandrs Slivkins. In Foundations and Trends in Machine Learning, 2019.