ICMLT 2023 Invited Speaker
Prof. Georg Frey, Saarland University, Germany
Georg Frey is full professor of Automation and Energy Systems with the Department of Natural Sciences and Technology at Saarland University, Saarbrücken, Germany since 2009. He is also scientific director of the Industrial Security group at Zentrum für Mechatronik und Automatisierung (ZeMA), Saarbrücken, Germany. Dr. Frey received the Diplom-Ingenieur (M.Sc.) degree in electrical engineering / control engineering from Karlsruhe Institute of Technology, Karlsruhe, Germany, in 1996 and the Doktor-Ingenieur (Ph.D.) degree in electrical engineering/automation from the University of Kaiserslautern-Landau, Kaiserslautern, Germany, in 2002. He was associate professor of Agent-Based Automation Systems with the University of Kaiserslautern, Visiting Professor at ENS Cachan, France, and researcher with the Smart Factory group at German Research Center for Artificial Intelligence (DFKI). He has been a co-chair and a member of several technical committees of IEEE and ifac. Dr. Frey has a strong background in control and automation science. He applies this in current research to the modeling and optimization of automation and energy systems. Similar technological trends in both domains like distributed networked structures, autonomy, intelligent control, and data based optimization as well as associated problems like complexity, reliability, safety, and security allow the application of similar methodological approaches and cross-domain learning. Component-based modeling and simulation is one of them, Machine Learning technologies are another. He is an IEEE Senior Member.
Speech Title: Machine Learning in Automation and Energy Systems Engineering
Engineering of automation and energy systems is mainly based on detailed mathematical modeling and analysis of the underlying physical (chemical, biological, societal, economic) systems. To this end a variety of tools is available. In recent years, due to the availability of more powerful tools and computational resources, in the design of complex systems analytical investigations are often supported or replaced by advanced (component based, distributed) simulation techniques and optimization schemes based on these (e.g.: Monte-Carlo Simulation, genetic approaches). Nowadays we see a similar trend in the application of Machine Learning techniques to support or replace physics-based modeling. In this talk, examples from the applications of Machine Learning techniques in the design auf automation and energy systems will be given. Based on these examples opportunities and risks will be identified and discussed, thus opening the discussion on future applications. Based on this discussion and further insights gained, the talk will finally focus on the question whether as simulation techniques did not make analytical approaches obsolete will the same hold true for Machine Learning techniques in respect to detailed mathematical system modeling. The answer on this question depends not only on the available Machine Learning techniques but also on the future education of engineers and data scientists. The availability of specialists with the right training is seen as a main obstacle for further successful application of Machine Learning technologies in classical automation and energy engineering domains.