With this event, we would like to get and provide an overview of activities concerning ML in and for SimTech. This will both involve successful applications of ML in different disciplines as well as presentations of ML techniques that are versatile and may be beneficial in various other simulation contexts. The event should serve as opportunity for networking, to bring together and ideally link ML method developers and application fields in mechanics, fluid dynamics, material science, chemistry, physics, etc.
The preliminary schedule is the following:
- Friedrich Solowjow & Sebastian Trimpe (MPI-IS): "Machine learning for dynamics and control"
- Hannes Eschmann (ITM): "Learning-Based Model Predictive Control using Gaussian Processes"
- Julian Hay & Jörg Fehr (ITM): "Surrogate Models for Crash Simulations"
- Johannes Kästner (ITC): "ML to approximate potential energy surfaces"
- Blazej Grabowski (IMW): “Momentum Tensor Potentials for thermodynamic simulations of complex materials with ab initio accuracy”
- Angel Diaz Carral & Maria Fyta (ICP): "Deep Learning Model for DNA Reads through Nanopores"
- Felix Fritzen (IAM): "Adaptive kernel regression and image based surrogate modeling in multiscale modeling"
- Jim Magiera & Christian Rohde (IANS): "Constraint-Aware Neural Networks for Riemann Problems"
- Marius Kurz & Andrea Beck (IAG): "Deep Residual Nets for Subgrid Model Learning in Large Eddy Simulation"
- Timothy Praditia & Wolfgang Nowak (IWS): "Physics-informed ANNs for dynamic, distributed and stochastic systems"
- Ingrid Blaschzyk & Ingo Steinwart (ISA): "Improved Classification Rates for Localized SVMs"
- Dirk Pflueger: "Sparse Grids for ML and ML for HPC"
- Tizian Wenzel & Gabriele Santin & Bernard Haasdonk (IANS): "Kernel Methods for Machine Learning and Surrogate Modelling"
If you are interested to follow these presentations you are welcome to join.