Machine Learning in der Technischen Mechanik

What it is about

You will learn how methods of Machine Learning can be applied to problems of numerical technical mechanics. The module conveys fundamentals, model reduction methods, as well as neural networks and shows how valid prediction models can be generated from data and used practically. In the end you can assess suitable ML methods and apply them to engineering problems.

What you will be able to do

• State and apply basic concepts, algorithms and functioning of Machine Learning
• Distinguish differences and peculiarities of various ML methods
• Assess applicability of the discussed methods for specific applications
• Carry out steps for model reduction
• Prepare datasets for ML algorithms
• Perform classification and feature extraction

What the module consists of

• VorlesungDelivery of theoretical foundations and demonstration on the data set; lecture materials and recordings are provided
• ÜbungPractical programming tasks in Python to apply what has been learned; solutions are discussed

Teaching method

• Vortrag / Präsentationto convey the theoretical foundations
• Tablet-PC mit Beamer / Tafelanschriebvisual representation of content and derivations
• Programmieren am Rechner (Python 3.x)practical implementation of ML methods and model building
• Kleinprojekteapplication to practical engineering problems, including handling of commercial FE software
• Bereitstellung von Materialien über TUM-Moodleaccess to lecture and exercise materials