Christoph Goebel studied computer science and economics at the Karlsruhe Institute of Technology, the École Polytechnique Fédérale de Lausanne, and Carnegie Mellon University. He received his doctorate from the Humboldt University in Berlin and then did research at the University of California in Berkeley and the Technical University of Munich, where he received his habilitation in 2016 from the Department of Computer Science. Until his appointment to the Professorship for Energy Management Technologies in October 2022, he worked in the private sector as part of innovation projects on machine learning applications. Christoph Goebel's research work focuses on the cross-sector development of new technologies for the intelligent control of energy systems. He applies the latest information technologies, in particular distributed systems, data-driven control algorithms and machine learning, for the optimal control of distributed energy resources such as solar systems, batteries, heat pumps, electric cars, thermal storage, and hydrogen systems.
How did you become who you are?
The support of my family and a very good basic education combined with great freedom for independent development certainly played a role that should not be underestimated. In addition, I had very good access to knowledge even before the Internet. For example, I remember having access to a shelf full of National Geographic magazines that I would regularly delve into as a child. I have always had a strong tendency to get to the bottom of things and to understand how technical systems, but also the economy, work. It was therefore difficult for me at the beginning of my scientific training to commit myself to one topic.
The only thing that became clear to me quickly was that I wanted to work creatively, i.e., really want to design and build technical systems. The reason I decided to study computer science and economics together was on the one hand my intensive occupation with programming at the time and on the other hand my first company, which I founded when I was still at school. I still follow a design-oriented and interdisciplinary approach in my research. Like many of my colleagues, I also believe that the next great breakthroughs in science will take place at the interfaces between disciplines.
I came to my research topic during my time as a postdoc at the University of California at Berkeley. The stay there was a stroke of luck for me, as it gave me the time and the best possible environment to search for the right research topic. On the one hand, the intelligent control of future energy systems seemed highly relevant to me at the time, on the other hand I also had the feeling that through my training and research I already had a lot of methodological knowledge for working on this topic. The contact to excellent scientists who were already active in this field did the rest. Since my time at Berkeley, I have also spent a lot of time in industry working on machine learning research and development projects. This field has made a leap forward, especially in recent years, and I am convinced that we can use the latest methods from this exciting area not only for new chatbots, but also for building a sustainable future.
What will be your first research project at TUM?
It is very important to me that our research has a positive effect on society. In our case, this is accelerating the development of our economy towards sustainability, especially in energy generation, distribution, and usage. Therefore, since I started in October, my colleagues and me have been working together on several prototype systems that demonstrate the potential of the latest information technologies in the energy transition and will ideally be used beyond scientific research, for example in cooperation with TUM Venture Labs.
With a first prototype, we will investigate how deep reinforcement learning can be used in energy management systems for buildings. This is a technology that enables purely data-driven control of very complex systems and was also used by Google Deepmind, for example, to achieve major breakthroughs in the field of AI. I believe that using deep reinforcement learning in the energy sector could allow us to achieve similar breakthroughs as in other areas.
A second prototype will be an operating system for energy management - you can think of it like a kind of Windows for the energy sector, on which you can easily install new software without having to buy new hardware. It is currently not possible to simultaneously use different energy management solutions with the same hardware, for example electric cars, energy storage devices or heat pumps. In addition, the development of innovative software often fails due to networking with the physical world and is very time-consuming since all software components always must be developed from scratch.
Finally, a third prototype addresses a problem that has been slowing down the development of innovative software systems in the energy sector for some time. Unfortunately, there is still no digital platform on which data, models, and algorithms for the development of energy management systems can be exchanged between researchers and further developed together. In other areas, such as image and language processing, this is already established practice, so why shouldn't it also work in the energy sector? Therefore, we would like to show what such a platform can look like and validate the corresponding functionality in a community-based approach.
What change do you hope to see in the future?
For various reasons, we in Germany face particularly big challenges in this energy transition. We have chosen a very complex and innovative approach and must be successful with it in a very short time. I hope that we, as a society, will accept this challenge together and see the great opportunities that come with finding a sustainable solution to our current energy problems. If we can develop and implement these technologies in Germany with our difficult starting position, then I see good chances that it will also work worldwide. I would like to be able to say to my children for once that I have done everything in my power to give them a good livelihood.
Especially in interdisciplinary research, you often don't get very far on your own: It's about learning from others and contributing your own knowledge as productively as possible in different research collaborations. In addition, I really enjoy working with other people. I therefore hope that I will find many colleagues in science, but also in industry, with whom I can work trustingly and effectively.
In addition to research on the topics already described, teaching is another central task. On the one hand, I try to structure my courses in such a way that students feel like delving into complex subject areas and learning new skills. On the other hand, I want to offer students a safe environment to try out their ideas, thereby encouraging them to trust themselves and their abilities. In my opinion, this cannot be done with traditional lectures alone, but requires new teaching formats - for example labs and internships in which students can build their own energy management system, for example. However, offering such courses requires a lot of resources on our side and I hope that I will still be able to gradually develop and offer them.