Table of contents

The course is connected to the following study programs

Teaching language

English

Prerequisites

Probability and Stochastic processes, Linear algebra, Optimization, Calculus, basic Signal Processing and Programming. Students can use different programming languages and libraries for the implementation of algorithms, such as: Python (e.g. TensorFlow, Pytorch) or Matlab libraries.

Recommended previous knowledge

IKT 724 Deep Learning and IKT 720 Optimization

Learning outcomes

Upon successful completion of this course, the students are expected to:

  • have a) knowledge of the main features of Deep Reinforcement Learning (DRL) that distinguishes it from other standard machine learning methods, and its connections with Optimization, b) knowledge to be able to correctly formulate a relevant application problem as a DRL problem, and justifiably identify the most suitable type of algorithm from the different possible classes of DRL algorithms, c) knowledge and understanding of the techniques to address the main challenges in DRL, as well as their strengths, limitations and properties, and the criteria (relevant metrics) for analysing and evaluating the performance of DRL algorithms.
  • be able to a) design, implement and simulate DRL algorithms to solve efficiently practical decision-making problems, b) evaluate experimentally the performance of DRL algorithms with respect to the relevant performance metrics.

Contents

Theoretical and algorithmic foundations of Deep Reinforcement Learning (DRL). Topics: Introduction and formal framework of DRL, Value-based algorithms, Deep Neural Networks in DRL, Policy-gradient algorithms, Actor-critic methods, Model-based algorithms, Other advanced topics (e.g. imitation learning, inverse reinforcement learning, multi-agent reinforcement learning, adversarial and safe reinforcement learning). Applications will be explored in the guided reading and/or exercises and final project implementation.

Teaching and learning methods

Lectures in the form of seminars and discussions with students, guided reading and/or exercises, student presentations, final project, self-study.

Examination requirements

Attendance to lectures, approved final project report.

Examinations

The final grade: Pass (A or B) or fail based on exercises and/or final project. The assessment criteria will be based on both technical correctness and clarity of exposition.

Student evaluation

The course will be evaluated in accordance with the quality system for education, chapter 4.1.

Offered as a free-standing course

Yes. Subject to availability or capacity.
Last updated from FS (Felles studentsystem) Oct. 15, 2026 8:28:16 AM