Deep Learning for PDEs in Engineering Physics

What it is about

In this internship you will learn how deep learning methods are used to solve partial differential equations (PDEs) in engineering applications. By the end you will be able to develop and apply neural networks with PyTorch to solve parametrized PDE problems, e.g. from fluid mechanics, materials science, or heat transfer.

What you will be able to do

• Knowledge of deep learning and neural networks
• Develop and train neural networks with PyTorch
• Apply deep learning methods to solve (parametrized) PDEs
• Implement deep learning algorithms for PDEs with PyTorch
• Apply deep learning methods to real engineering-physical PDE problems

What the module consists of

• PraktikumTheoretical basics at the beginning of each session (~45 min) and practical exercises to deepen understanding as well as development of applications in PyTorch
• SelbststudiumHomework to deepen the contents
• TutoriumTutors are available to answer questions

Teaching method

• Kurzvortrag mit FolienConveying the theoretical foundations at the beginning of each session
• Computergestützte IllustrationenIllustration of concepts and methods
• Praktische Übungen in PyTorchApplication and implementation of the methods on practice examples
• HausaufgabenIndependent deepening and consolidation of the material