Reinforcement Learning Practical

Faculteit Science and Engineering
Jaar 2021/22
Vakcode WBAI015-05
Vaknaam Reinforcement Learning Practical
Niveau(s) bachelor
Voertaal Engels
Periode semester I b

Uitgebreide vaknaam Reinforcement Learning Practical
Leerdoelen This course aims to introduce the student to most of the concepts underlying the field of Reinforcement Learning, with a particular focus on model-free Reinforcement Learning.
At the end of the course, the student will be familiar with:
- The mathematical framework used in Reinforcement Learning (lecture 1)
- The exploration-exploitation dilemma (lecture 2)
- The development of Dynamic Programming algorithms (lecture 3)
- The creation of tabular based model-free algorithms (lecture 4)
- The use of function approximators (lecture 5)
- The main challenges of Reinforcement Learning (lecture 6)
Omschrijving Reinforcement Learning (RL) is the branch of machine learning that aims to teach agents how to interact with an environment through trial and error. Such interaction is usually modeled as a Markov Decision Process where the end goal of the agent, sometimes called the learner, is that of maximizing a certain reward signal. Unlike other machine learning approaches, such as the arguably more popular supervised learning one, RL is largely considered more challenging as an agent is deprived of any external supervision. Therefore, it can only rely on its own personal experience while learning. In this course, we will see how one can train such agents by characterizing RL algorithms both from a theoretical perspective as well as from a more practical one. To this end, six theoretical lectures will be given, the content of which will have to be, in part, put into practice in two different assignments and in a final project. Theoretical lectures will be given every Monday from 9:00 to 11:00, which will then be followed by a computer practicum from 15:00 to 17:00.
Uren per week
Onderwijsvorm Hoorcollege (LC), Practisch werk (PRC)
Toetsvorm Opdracht (AST), Verslag (R)
(The final grade is based on the grades obtained for: i) Assignment 1 (25%): which consists in a coding assignment related to lecture 2; ii) Assignment 2 (25%): where students have to solve some simple mathematical problems related to lecture 3 and lecture 4; iii) Final Project (50%): a RL project of the student's choice. Students are allowed to work alone or in groups of a maximum of two people.)
Vaksoort bachelor
Coördinator M. Sabatelli, MSc.
Docent(en) N. Orzan, MSc. , M. Sabatelli, MSc.
Verplichte literatuur
Titel Auteur ISBN Prijs
Reinforcement Learning: An Introduction 2nd 2018 Sutton. and A. Barto
Entreevoorwaarden Mandatory: Autonomous Systems (WBAI002-05), Imperative Programming (WBAI003-05).

If the mandatory requirements are not met, only the Board of Examiners of the AI BSc may grant an exemption. Exchange students are assumed to have gone through this through their Learning Agreement; pre-master's students through the Board of Admissions - other external students are judged case-by-case.
Opmerkingen This course unit has a capacity limit. More information about capacity-limit courses can be found here. This course has an intended limit of 80 participants.
Artificial Intelligence (BSc) is a Fixed Quota (Numerus Fixus) programme. As a consequence, their courses (course code WBAI) are closed for students that are not registered under the AI BSc programme, unless the course is part of the mandatory curriculum of their programme. If you wish to take this course in your minor – or as part of a so-called ‘unofficial’ pre-master’s – please use the official procedure through the Board of Examiners form.
Opgenomen in
Opleiding Jaar Periode Type
BSc Artificial Intelligence 3 semester I b keuzegroep