| Lecturer (assistant) | |
|---|---|
| Language of instruction | German |
| Dates | See TUMonline |
Content
In today's engineering world, numerical methods are increasingly used to shift the development process of products and machines step by step into virtual space. The resulting models reflect an image of reality based on fundamental laws of physics. However, the calculation times for complex systems can significantly exceed the benefits and thus lose their attractiveness. An alternative approach to this is the application of machine learning. Therefore, the module gives an introduction to Machine Learning (ML) and discusses the theory of optimization in ML, forms of modeling in ML, model reduction methods, principal component analysis, singular value decomposition, Dynamic Mode Decomposition (DMD), Artificial Neural Networks, supervised ANN, deep-learning, Bayesian deep-learning as well as applications.
- ... reproduce the basic concepts, algorithms, and operation of Machine Learning and apply them appropriately
- ... differentiate the peculiarities of different machine learning methods
- ... identify the essential characteristics of the discussed methods concerning specific applications in order to assess their applicability, consequently
- ... perform appropriate steps for model reduction
- ... prepare data sets to apply appropriate machine learning algorithms
- ... perform classifications and feature extraction
Organizational Matters
Lecture:
- Weekly regular class, 2 SWH duration
- Presentation of theoretical contents and recordings of the essential content as videos
Exercise:
- Weekly regular class, 2 SWH duration
- Practical tasks to learn the application of machine learning to a data set
- Self-reliant execution of exercises, students implement their code under supervision
- Usage of Python
Examination:
- Written (90min) or oral (30min) open-book exam (decision in the respective semester)
Recommended Previuous Knowledge
- Engineering Mechanics
- Mathematics and statistics (B. Sc. Level)
The above mentioned recommended previous knowledge is not obligatory to successfully complete this course. If you are indecisive about your participation in this course, please contact the responsible course assistant.