Details

KoMSO Academy 2021

From 14-16 September, the first KoMSO Academy "Combining model- and data-based approaches for industrial problems", organised by the Centre for Industrial Mathematics (ZeTeM) and the Committee for Mathematical Modeling, Simulation, and Optimization (KoMSO), took place in Bremen in hybrid form.

The KoMSO Academy 2021 combined a challenge workshop with renowned speakers from industry and science on the topics of energy and mobility with a training and hands-on session on the topic of deep learning.

In the keynote lecture of the Challenge Workshop, Prof. Wil Schilders from the University of Eindhoven gave a far-reaching insight into real and artificial intelligence (AI) for science and engineering. Afterwards, the participants had the chance to learn how Siemens, Deutsche Bahn and EWE integrate the challenge and opportunity of artificial intelligence into their research and development processes. In addition, the latest scientific results on stochastic and data-based models for electricity demand and on modelling of turbulent flows by generative learning were presented in two scientific lectures.

In the training course, an introduction to training neural networks with PyTorch was given, model-based classical approaches (ISTA) and data-based methods (LISTA) as well as their combination were covered. The programming tasks for the participants, who were mainly connected online, came from a real application scenario from computed tomography.

We would like to thank all speakers for their contributions here in Bremen and hope that all participants, whether in person or online, took away new ideas and suggestions from the event.

You can find more information here.

 

AI industrial problems