This course will take place as an online event. The link to the streaming platform will be provided to the accepted registrants only.
This training course will highlight interactive analysis for ML/AI applications in the research domain of combustion theory. The program will focus on fluid mechanics, fundamentals and current challenges in combustion, as well as the use of Machine Learning (ML) and High-Performance Computing (HPC) to approach simulations of turbulent reacting flows.
The course is offered in cooperation with the Centre of Excellence for Combustion (CoEC).
Preliminary Agenda:
|
Time |
Title |
Presenter |
|---|---|---|
|
08:30–09:15 |
Welcome and introduction to interactive supercomputing for machine learning on HPC |
Jens Henrik Göbbert (Jülich Supercomputing Centre) |
|
09:15–10:00 |
Introduction to machine learning: artificial and convolutional neural networks |
Ludovico Nista (RWTH Aachen University) |
|
10:45–11:15 |
Coffee break |
|
|
11:15–12:00 |
Hands-on session: - linear regression and rain prediction via ANN, a digit classifier using CNN - a flamelet representation with ANN |
Julian Bissantz (TU Darmstadt) |
|
12:00–13:30 |
Lunch break |
|
|
13:30–14:15 |
Machine learning for combustion modeling: GAN modeling of sub-filter turbulence |
Dr. Temistocle Grenga (RWTH Aachen University) |
|
14:15–15:00 |
HPC with Machine Learning: multi-node multi-GPU training: theory and applications |
Ludovico Nista (RWTH Aachen University) |
|
15:00–15:30 |
Coffee break |
|
|
15:30–16:15 |
HPC and ML for Combustion Simulation: Embedding ML into CFD code |
Dr. Peicho Petkov (NCSA) |
|
16:15–17:00 |
Advanced applications of AI super-resolution to combustion |
Mr. Mathis Bode (Jülich Supercomputing Centre) |
Prerequisites:
Some knowledge of Computational Fluid Dynamics
Date:
8 December 2022, 08:30-17:00
Application
Registrations are only considered until 1 December 2022, the maximal number of participants is limited. Applicants will be notified, whether they are accepted for participitation.
Instructors:
Employees from Jülich Supercomputing Centre (JSC), RWTH Aachen University, TU Darmstadt University, and NCSA.
