Description
We approach data science via a modern optimization lens. In this course, we cover mathematical optimization methods with applications in data science. The emphasis is on modeling of convex optimization problems, comprehension of algorithms for solving them and using a dedicated software.
Convex optimisation constitutes an important class of problems in data science. Their computational properties make them attractive to be applied on a wide range of real-world problems in data science and business analytics.
Themes covered
1. Convex sets and functions
2. Canonical problem types
3. Applications of convex optimization
4. Optimality conditions and duality
5. Gradient descent
6. Subgradient method
7. Stochastic gradient descent
8. Newton and Quasi-Newton methods
9. Coordinate descent
10. Frank-Wolfe method
Important notes
Course in French : MATH 60608
