Reinforcement Learning (DAT605)
This course introduces AI and optimization through hands-on learning! This course makes tackling optimization challenges easy and fun, showing you how to streamline complex models, build smart systems that learn and make decisions on their own, and uncover patterns in data to guide better outcomes. Gain skills that apply to various fields, from efficient model design to real-world problem-solving.
Course description for study year 2025-2026. Please note that changes may occur.
Course code
DAT605
Version
1
Credits (ECTS)
5
Semester tution start
Autumn
Number of semesters
1
Exam semester
Autumn
Language of instruction
English
Content
NB! This is an elective course and may be canceled if fewer than 10 students are enrolled by August 20th.
We bridge the gap between supervised and unsupervised learning. While supervised learning relies on labeled data and unsupervised learning seeks hidden patterns without labels, Reinforcement Learning (RL) takes a unique approach. It combines learning from examples with decision-making in complex environments—learning from experience to make optimal choices over time.
In this course, you'll explore the essentials of RL, from function approximation to advanced decision-making. RL draws from computer science, psychology, and neuroscience, making it a powerful tool for a wide range of applications, from gaming, game theory and control systems to swarm intelligence and optimization.
Learning outcome
This course introduces the essential theory and hands-on programming skills in reinforcement learning (RL) in an engaging and accessible way.
Knowledge
- Introduction to Reinforcement Learning : Get a solid grounding in RL and see how it applies to real-world decision-making.
- Resources and Setup : Familiarize yourself with course materials and coding tools. A crash course on neural networks and deep learning frameworks is also included.
- Core Methods : Discover how agents learn from their choices and feedback to make smarter decisions.
- Learning from Experience : Understand how agents adapt and improve based on their actions and the rewards they receive.
- Advanced Techniques : Explore methods that help agents make effective decisions, even in complex situations.
Skills
- Programming skills
- Basics of algebra, probability, and statistics
- Familiarity with Python, Numpy, and Matplotlib
General Competence
Gain a foundational understanding of neural networks, deep learning, and CNNs.
This course makes reinforcement learning approachable, equipping you with the tools and insights to apply these skills to real-world challenges and innovations.
Required prerequisite knowledge
Recommended prerequisites
Exam
Project report and oral presentastion
| Form of assessment | Weight | Duration | Marks | Aid |
|---|---|---|---|---|
| Project report (group) | 1/2 | Letter grades | ||
| Oral exam | 1/2 | Letter grades | None permitted |
Two seperate assessments are conducted for the course completion and grades: Group Project and Indivudal Oral Exam- The project will be completed in assigned groups, which will be announced at the start of the project period. The project assessment, including a report with source code, accounts for 50% of the final grade. Students will have three weeks to complete the project, and all group members will receive the same grade.- In addition, each student will have a 30-minute individual oral exam, contributing the remaining 50% of the grade. This exam will assess each student’s understanding and application of the course material.If a student does not pass either the project or the oral exam, they will need to retake the relevant component(s) when the course is next offered.Any absences due to illness or other reasons should be communicated to the lecturer and the exam office as soon as possible.