Checking date: 16/04/2024


Course: 2024/2025

Optimization
(19279)
Master in Machine Learning for Health (Plan: 480 - Estudio: 359)
EPI


Coordinating teacher: VAZQUEZ VILAR, GONZALO

Department assigned to the subject: Signal and Communications Theory Department

Type: Electives
ECTS Credits: 6.0 ECTS

Course:
Semester:




Requirements (Subjects that are assumed to be known)
Students are expected to have a solid background in - Linear Algebra Prior knowledge on optimization is not required.
Objectives
Optimization theory is nowadays a well-developed area, both in the theoretical and practical aspects. This graduate course introduces the basic concepts for solving optimization problems and illustrates this theory with many recent applications in signal processing and machine learning. Students attending this course will: - Develop a solid theoretical basis for solving convex optimization problems arising in industry and research. - Learn manipulations to unveil the hidden convexity of optimization problems and relaxation techniques to treat non-convex optimization problems. - Be able to characterize the solution of convex and non-convex optimization problems either analytically or algorithmically. - Learn the usage of some of the more popular optimization toolboxes.
Skills and learning outcomes
Description of contents: programme
Unit 1. Introduction - Optimization problems and constraints - On closed-form optimization: analytical versus algorithmic solutions - Types of optimization problems - Modelling and applied linear algebra Unit 2. Convex Optimization - Convex sets and convex functions - Convex optimization problems - Disciplined convex programming, CVX - Quadratic optimization - Lagrange duality and KKT conditions Unit 3. Optimization Algorithms - Local optimization algorithms - Stochastic optimization - Global optimization - Integer programming and metaheuristics Unit 4. Applications - Optimization for machine learning
Learning activities and methodology
- Theoretical sessions: theoretical basis of optimization theory, illustrated with different applications and examples. Material for out-of-class work. - Problem sessions: formulation and solution of exercises motivated by different problems from communications, signal processing and machine learning. - Practical sessions: popular toolboxes for convex and non-convex optimization. The proposed projects will be solved in Matlab and/or Python programming environments.
Assessment System
  • % end-of-term-examination 0
  • % of continuous assessment (assigments, laboratory, practicals...) 100

Calendar of Continuous assessment


Basic Bibliography
  • S. Boyd and L. Vandenberghe. Convex Optimization. Cambridge University Press. 2004
Recursos electrónicosElectronic Resources *
Additional Bibliography
  • A. Zhang, Z. C. Lipton, M. Li and A. J. Smola. Dive into Deep Learning. Online interactive book: https://d2l.ai. 2019
  • S. Boyd and L. Vandenberghe. Introduction to Applied Linear Algebra - Vectors, Matrices, and Least Squares. Cambridge University Press. 2018
Recursos electrónicosElectronic Resources *
  • A. Zhang, Z.C. Lipton, M. Li and A.J. Smola · Dive into Deep Learning : https://d2l.ai
  • S. Boyd and L. Vandenberghe · Introduction to Applied Linear Algebra - Vectors, Matrices, and Least Squares : http://vmls-book.stanford.edu/
(*) Access to some electronic resources may be restricted to members of the university community and require validation through Campus Global. If you try to connect from outside of the University you will need to set up a VPN


The course syllabus may change due academic events or other reasons.