Annual Workshop Speaker: Frank Dressel
How to trust digital twins
Frank Dressel, Acting Head of Department Dpt. Software Methods, Institute of Software Methods for Product Virtualization, German Aerospace Center (DLR)
ABSTRACT
Digital twins are models in the virtual space, which may be accompanied by physical objects. A digital twin provides access to data which is related to the underlying model. The data may come from data storage, other virtual models, simulations, measurements or other sources which are relevant for the digital twin. In many cases, the digital twin is more than a simple data storage and contains functional modules like methods of machine learning for predictive values, continuous data integration or federated components to allow accurate predictions about the (future) state of the model. For a user of such a digital twin, the data often looks like generated from a black box. He can only believe in the data but cannot use it in an informed way. To help him make an informed decision and assess the data, further meta information like provenance or data lineage is needed. Dr. Dressel will sketch the basic ingredients for a digital twin with provenance, show the basic building blocks and highlight questions which need to be answered to be able to trust data from digital twins.
ABOUT THE SPEAKER
After his studies in physics, Dr. Dressel made his PhD at the Technische Universität Dresden in the field of computational physics. During his postdoc he worked in the field of bioinformatics, gave lectures and worked as freelancer for a spin-off. In 2011 he moved to industry and worked for different companies in the field of software engineering with a special focus on software development, data management, data science, technology scouting and prototyping. In 2019, Dr. Dressel joined the German Aerospace Center (DLR) and currently works as acting head of the department Software Methods at the Institute of Software Methods for Product Virtualization in Dresden.