Practical information
The tutorial takes place at the J. A. Dieudonne laboratory on the Valrose campus.
Address: Parc Valrose, 28 Avenue Valrose, 06000 Nice, map.
Conference page: mode2026.sciencesconf.org
Lecturers
Outline
Schedule and course material
The course is organized across three main teaching blocks. Lab notebooks and solutions open directly in Colab.
Monday afternoon
Session 1 - Intro and Part 1
Introduction to the tutorial, followed by the first lecture blocks and hands-on material.
Lab 1 solution: 🐍 open solution in Colab .
Tuesday morning
Session 2 - Part 2
Continuation of Part 1 with the second slide deck and the next lab session.
Solution: 🐍 session 2 in Colab.
Tuesday afternoon
Part 3
Algorithm design material, with a dedicated starter notebook for the practical session.
Solution: 🐍 session 4 in Colab .
Challenge question / Feedback
Submit your most challenging question to the others
Open the form
Join the quizz!
🧠 Wooclap quizz!
🧠 Quizz - at your own pace
GitHub projects
If you enjoyed the material, starring these repositories is an easy way to support the open-source tools and course creators behind the tutorial.
This kind of support is also important for the visibility and recognition of the projects.
Context and references
Performance estimation problems were introduced in [1]. This mini-course mainly follows the framework and developments from [2-4], together with the informal introduction in [5].
The labs are based on the PEPit code package for performance estimation problems [4].
- [1] Performance of first-order methods for smooth convex minimization: a novel approach
- [2] Smooth strongly convex interpolation and exact worst-case performance of first-order methods
- [3] Exact worst-case performance of first-order methods for composite convex optimization
- [4] B. Goujaud, C. Moucer, F. Glineur, J. M. Hendrickx, A. B. Taylor, and A. Dieuleveut. PEPit: computer-assisted worst-case analyses of first-order optimization methods in Python Mathematical Programming Computation 16(3), 337-367.
- [5] Computer-aided analyses
Acknowledgments
Funding
We thank Daniel Berg Thomsen for feedback on the content of this mini-course.
We also thank Damine Ngo for his work on building the tutorial UI.
The project was co-funded by the European Research Council and by the Agence Nationale de la Recherche.
The project benefitted from the support of Hi! Paris, in particular through the engineering team (Damien Ngo).
