The subject area of machine intelligence and, in particular, machine learning, taking into account real challenges of complex application domains, is a rapidly expanding field of knowledge and the subject of numerous research and development projects.

The lecture "Machine Learning 2" deals with modern advanced methods of machine learning such as semi-supervised, self-supervised and active learning, deep neural networks (deep learning, CNNs, GANs, diffusion models, transformer, adversarial attacks) and hierarchical approaches, e.g. reinforcement learning. Another focus is the embedding and application of machine learning methods in real systems.

The lecture introduces the latest basic principles as well as extended basic structures and elucidates previously developed algorithms. The structure and the mode of operation of the methods and methods are presented and explained by means of some application scenarios, especially in the field of technical (sub) autonomous systems (vehicles, robotics, neurorobotics, image processing, etc.).

Learning objectives:

• Students understand extended concepts of machine learning and their possible applications.
• Students can classify, formally describe and evaluate methods of machine learning.
• In detail, methods of machine learning can be embedded and applied in complex decision and inference systems.
• Students can use their knowledge to select suitable models and methods of machine learning for existing problems in the field of machine intelligence.

Recommendations:

Attending the lecture Machine Learning 1 or a comparable lecture is very helpful in understanding this lecture.

Die Foliensätze sind als PDF verfügbar

Weiterführende Literatur

• Deep Learning - Ian Goodfellow
• Artificial Intelligence: A Modern Approach - Peter Norvig and Stuart J. Russell
• Machine Learning - Tom Mitchell
• Pattern Recognition and Machine Learning - Christopher M. Bishop
• Reinforcement Learning: An Introduction - Richard S. Sutton and Andrew G. Barto
• Deep Learning - Ian Goodfellow, Yoshua Bengio, Aaron Courville

Weitere (spezifische) Literatur zu einzelnen Themen wird in der Vorlesung angegeben.