Optimization problems need to be solved in almost every domain. Nurses and doctors, for example, need to be assigned to shifts in a hospital without violating given constraints, investment portfolios have to be built in order to maximize the return, or the many parameters of a weather forecast model have to be chosen in order to best fit previously measured data. Machine learning is no exception: optimization problems occur, for example, in reinforcement learning or during the hyperparameter tuning of machine learning algorithms.
This introductory course on optimization for machine learning aims at teaching the basic knowledge about optimization theory and the design and analysis of optimization algorithms. One goal is to provide the necessary background that will allow the participants to practically address the various optimization problems they might encounter in the future.
The lecture covers both fundamental aspects in discrete optimization (such as greedy algorithms, dynamic programming, heuristics) and continuous optimization (introducing gradient-based methods as well as derivative-free algorithms). If time allows, we will also cover more specific aspects such as multiobjective optimization.
The course will be entirely taught online on Thursday afternoon from 1:30pm till 4:30pm using Zoom. Office hours will take place from 4:30 till 5pm.
Before the 4th of November, please send an email to Dimo Brockhoff with title "Register for TC2 Optimization class" to register your email address. You will then be sent a link to be part of the whatsapp group of the class. Links for attending the class via zoom will be sent via this whatsapp group.
What do I need to attend the class?
You need a computer where you can use Zoom/MS Teams and a good access to internet, some pens and some paper to take notes of the class and work on the exercices and examples, a mobile phone with whatsapp and camera where you can take pictures of your solutions of the exercices (and quickly share them).
The first lecture will be broadcast via MS Teams. You can join by following this link.
The lecture is given by Anne Auger and Dimo Brockhoff from November till December 2020.
| Anne Auger and Dimo Brockhoff | ||||||
|