Probabilistische Ersatzmodellierung

What it is about

You will become familiar with and apply various procedures of probabilistic surrogate modeling. In the end you can design surrogate models, assess their accuracy and use them for uncertainty propagation, sensitivity analysis and reliability analysis, and implement selected methods in Matlab or Python.

What you will be able to do

• Understanding the entire surrogate modeling process (experimental design, calculation, validation)
• Knowledge of different surrogate approaches: regression, polynomial chaos, Gaussian processes, SVM, neural networks
• Evaluation and improvement of the accuracy of surrogate models
• Assessment of strengths and weaknesses of the methods for different UQ tasks
• Implementation of the methods in Matlab or Python

What the module consists of

• VorlesungDelivery of theoretical foundations and demonstration of code examples
• Übungen / AufgabenblätterApplication and deepening of the material; regular problem sets (biweekly)

Teaching method

• TafelvorlesungExplanation of theoretical concepts and derivations at an appropriate pace
• FoliensätzeIllustration of applications and structuring of content
• Code‑DemonstrationenShow how theory is implemented in Matlab/Python