This section provides an overview of the specialisation courses of our Micro Degree Programmes. Further details regarding locations, timetables, and the semesters in which the courses are offered can be found in the university’s online course catalogue.
AutoML - Automated Machine Learning
Responsible: Prof. Dr. rer. nat. Marius Lindauer
Institute of Artificial Intelligence
Description: Students will learn the fundamental principles of machine learning, including both classical machine learning and deep learning approaches. Upon successful completion of the course, they will be able to explain methods for hyperparameter optimisation and neural architecture search, and apply them to novel problem settings and datasets. In particular, they will be equipped to implement these techniques in practice in order to systematically improve the performance of machine learning algorithms on, for example, tabular or image data.
The syllabus includes the following topics: design spaces in machine learning, experimentation and visualisation, hyperparameter optimisation (HPO), Bayesian optimisation, additional black-box techniques, acceleration of HPO through multi-fidelity optimisation, architecture search I and II, dynamic approaches, and Beyond AutoML: algorithm configuration.
Credit Points: 5
Language: English
Computer Vision
Responsible: Prof. Dr.-Ing. Bodo Rosenhahn
Institute for Information Processing
Description: Computer Vision (also known as Machine Vision) refers to the algorithmic solution of tasks that are based on the capabilities of the human visual system.
The course Computer Vision serves as an interface with the modules Digital Signal Processing, Digital Image Processing, Machine Learning, and Computational Scene Analysis, and covers advanced methods of image analysis.
The topics covered include, among others, segmentation algorithms (e.g. active contours, graph cut), feature extraction, optical flow, as well as Markov Chain Monte Carlo methods (e.g. particle filters, simulated annealing).
In addition, participants are provided with a comprehensive overview of the research field.
The syllabus includes, among other topics: the Hough transform, keypoint detection, segmentation, optical flow, matching, and Markov Chain Monte Carlo techniques.
Credit Points: 5
Language: German
Deep Learning Foundations
Responsible: Dr. Sandipan Sikdar
Institute of Data Science
Description: Students will acquire a solid understanding of the fundamentals of deep learning by mastering modelling, training, and optimisation techniques applied to text, image, and graph data.
The syllabus includes, among other topics: the fundamentals of neural networks; training, optimisation, and regularisation in deep learning; convolutional neural networks (CNNs); recurrent neural networks; deep generative models; representation learning in text and graph domains; and applications such as image captioning and question answering.
The course covers both the theoretical foundations and practical implementation aspects of deep neural networks across a wide range of application domains. A particular emphasis is placed on successful use cases of deep learning and on the role of rich representations in enabling and enhancing these applications.
Credit Points: 5
Language: English
Reinforcement Learning
Responsible: Prof. Dr. rer. nat. Marius Lindauer
Institute of Artificial Intelligence
Description: In recent years, Reinforcement Learning (RL) has achieved some of the most remarkable advances in the field of Machine Learning (ML) – particularly in games (such as Go) and in robotics (e.g. RoboCup or autonomously navigating robots). Modelling the learning system as an agent acting within an environment enables learning through trial and error, allowing for conclusions that go beyond human expert knowledge.
RL is a rapidly evolving research area, with new algorithms and applications emerging continuously. This course begins by introducing the mathematical foundations of Reinforcement Learning and providing an overview of both its historical development and current state of research.
Upon successful completion of the course, students will be able to explain the theoretical foundations of core RL approaches and understand the current landscape of RL research. In the accompanying practical sessions, they will acquire hands-on skills in implementing various RL algorithms and gain familiarity with the typical RL pipeline – including learning environments, agent evaluation, and hyperparameter configuration.
At the end of the course, students will apply their acquired knowledge independently in an RL project of their own choice.
The syllabus includes, among other topics: Markov decision processes, value function approximation, policy search, model-based reinforcement learning, deep RL, and meta-RL.
Credit Points: 5
Language: English
Knowledge Engineering and Semantic Web
Responsible: Prof. Dr. Sören Auer
Institute of Data Science
Description: Participants will acquire a foundational understanding of Knowledge Engineering, including ontologies, knowledge graphs, reasoning, and inference. In addition, they will gain both theoretical and practical knowledge in working with established W3C standards for data exchange (RDF, SPARQL, RDFa, Microdata) as well as key technologies of the Semantic Web.
The aim of the course is to develop the ability to understand, analyse, and independently design knowledge models and ontologies.
The syllabus includes, among other topics: knowledge representation using RDF, ontology modelling with RDF Schema (RDFS) and OWL, and the SPARQL query language.
Credit Points: 5
Language: English
Visual Analytics
Responsible: Prof. Dr. Ralph Ewerth
Institute of Data Science
Description: Visual Analytics is concerned with the analysis, processing, and interactive visual representation of large and complex datasets, with the aim of uncovering new information and gaining insights from the data.
Within the framework of this course, students acquire foundational knowledge in data preprocessing as well as in the principles of human visual perception. They are introduced to a variety of visualisation techniques for different data types – including spatial, geographic, and hierarchical data – and are able to critically assess their respective advantages and limitations.
In addition, students explore key concepts and techniques of interaction and develop an understanding of how to design effective visualisations. By the end of the course, they are able to evaluate visualisation approaches in terms of expressiveness, interactivity, and technical feasibility, and to independently develop interactive visualisation systems using appropriate software libraries.
Syllabus / Topics covered:
- Introduction to interactive data and information visualisation
- Data types and fundamental processing steps
- Human perception and information processing
- Visualisation of spatial and geographic data
- Visualisation of trees, graphs, and networks
- Visualisation of texts, documents, and multimedia data
- Interaction: concepts and techniques
- Design of effective visualisations
- Comparison and evaluation of visualisation techniques and systems
Credit Points: 5
Language: German
