Latest

---

---

Instructors

Clément Royer
clement.royer@lamsade.dauphine.fr

Back to the general teaching page

Optimization for Data Science

M2 MIAGE ID/ID Apprentissage, Université Paris Dauphine-PSL, 2025-2026

---

Aim of the course

     Provide modern algorithmic tools for data science problems.

Course material

     Lecture notes (Jan. 19) PDF

Session 1: Basics of optimization

     Virtual board

Session 2: Gradient methods (practice)

     Notebook: introduction to gradient descent (without solutions)
     Notebook: introduction to gradient descent (with solutions)

Session 3: Gradient methods (theory)

     Virtual board
     Illustration notebook (accelerated methods, nonconvex problems)

Session 4: Stochastic gradient (theory)

     Virtual board

Session 5: Stochastic gradient (practice)

     Notebook on stochastic gradient (without solutions)
     Notebook on stochastic gradient (with solutions)

Session 6: Regularization

     Virtual board
     Illustration notebook (l2 and l1 regularization)

Session 7: Summary and exercises

     Virtual board

Tutorial resources

     Tutorial 1 (basics on optimization, with solutions) PDF
     Tutorial 2 (gradient descent, with solutions) PDF
     Tutorial 3 (stochastic gradient, with solutions) PDF
     Tutorial 4 (regularization, with solutions) PDF
     Tutorial 5 (exam 2024-2025, with solutions) PDF