Dr. Marco Breiling – “We can’t wait for the rest of the world”

March 21, 2022 | Portrait of one of our experts in neuromorphic hardware

Dr. Marco Breiling (51) wants to assist in shaping the future of Germany. The electrical engineer from Erlangen helps get successful high-tech products made in Germany. Inspired by the energy efficiency of a bee’s brain, his team develops clever microchips that process very large data volumes in an energy-saving manner. This achievement impressed the jury of the pilot innovation competition “Energy-Efficient AI System” organized by the German Federal Ministry of Education and Research (BMBF). Here, Marco Breiling gives fascinating insights into his field of research.

We have set ourselves a tough challenge. Although science doesn’t fully understand the nervous system of humans or even of simple animals, our current imperfect knowledge is the foundation upon which we seek to, say, make chips as efficient as the brain of a bee. This is something we haven’t yet achieved. Our chips are still a long way off from the energy efficiency of a bee’s brain, never mind brains as complex as those of other animals. But it’s precisely this challenge that makes our work so interesting!

Thinking big, making the supposedly impossible possible: we dare to try! Take humanity’s dream of flight: I’m motivated by the thought that it was birds that inspired people. Initial attempts had little success, but then people discovered further non-biological mechanisms – and today aircraft are much faster than birds and can carry much more weight.

Making progress toward an ambitious goal – this has often been my personal motivation. For my twelfth birthday, I asked for an electronics kit. I had read that with some add-ons to the basic kit, you could build your own television. But the project didn’t come to fruition: there were just too many expensive extra kits required. When I had expanded my electronics kits for a few years, personal computers came on the market, and soon I was putting my money there. But I still pursued my interest in electronics, just not by buying new sets. Instead, I played around with the existing kits, took part in the “Schüler experimentieren” and “Jugend forscht” contests for children and youth interested in STEM subjects, recreated circuits from electronics magazines, and later took part in the “Bundeswettbewerb Informatik” IT contest.

This was what inspired my interest in electrical engineering, which I went on to study in Karlsruhe. What motivated me back then was the prospect of a few semesters abroad, and I took full advantage of the opportunity. I studied in Trondheim, Paris and Southampton, all of which were inspiring and enriching experiences. I then went to Erlangen for my doctorate.

Our contribution should be to put Germany in the top division.

Next, I got an offer from the Fraunhofer Institute for Integrated Circuits IIS. At the time, I was interested in telecommunications; it was all about sending and receiving bits and bytes. This is important for mobile telephony and for satellite communications – even back then, I was working on algorithms and how to implement them on a chip. Today it’s artificial intelligence, and we’re working on ways to solve very complex tasks using machines.

Is it personal ambition that drives me? Well, I’m certainly the kind of person who likes to complete tasks they’ve set themselves. But it is curiosity above all that drives me. There are so many topics I’m interested in! My area of specialization is currently neuromorphic hardware, which is based on specialized computer architectures that fundamentally mirror the structures of neural networks. This involves concepts and devices or chips that are inspired by biology, or more specifically the brain, and are designed to carry out the calculations and perhaps one day even the training of neural networks.

The foundation of our work is the research being done at universities. At Fraunhofer IIS, we do pioneering work in implementation. Our contribution should be to put Germany in the top division when it comes to the development and use of neuromorphic hardware. A good example is the task we solved for the innovation competition run by the German Federal Ministry of Education and Research (BMBF). The question was: “What kind of chip is capable of detecting cardiac arrythmias and atrial fibrillation in ECG data with an accuracy of at least 90 percent while consuming the least amount of energy?” Our answer was “low-power, low-memory, low-cost ECG signal analysis with ML algorithms (Lo3-ML).” The chip contains non-volatile memories, in this case RRAMs, along with ultra-low-power circuits for reading and writing. This system records data while the AI algorithm is inactive. When data needs to be processed, the algorithm is activated very quickly to execute its task in extremely short order. As a result, the chip uses as much as 95 percent less energy than systems that are constantly active.

Now, if that was all a little bit abstract, maybe the image of a data highway will help: in a conventional computer, a broad route transports the data, but we create many very small, highly branched roads that are connected with each other. To put it in very general terms: on one chip, we can evaluate huge volumes of complex information in many tiny “mini-computers” collectively in a very energy-efficient manner.

We are exchanging information and expertise across discipline boundaries – and learning from each other.

To make this work, we’re working across discipline boundaries. This is an aspect that I’d like to emphasize to young people in particular. In my observation, young people have no specific idea of what scientific work actually is. It was that way for me, too: I was a country boy, and there were no academic role models in my family. I had no real idea of what university would be like, never mind what awaited me in a career after that.

We work at the intersection of engineering, mathematics, IT, physics and neuroscience. And in the case of Lo3-ML, we even work with non-technical professions. Everyone contributes their expertise and we exchange ideas and viewpoints and learn from each other. From time to time, we end up talking about non-technical matters. We like discussing philosophical, political and economic subjects. For my own part, I’m interested in general mechanisms, including economic and social ones; for example, how dictators manage to stay in power.

Or we discuss the question as to whether technology alone helps in addressing the great challenges facing humanity, such as climate change. As an engineer, I’m confident that we will help protect resources, for example, by means of technical innovations. However, I see it as a long-term undertaking. If we continue to consume products, meat, travel and all these things at the same rate as now, and then we add in the developing and emerging countries doing the same, then technology won’t be able to make much of an impact here. Cattle will suffer, and so will the environment. In the medium term, we in industrialized countries undoubtedly need to learn to be a bit more frugal at times, while also offering greater support to developing nations. 

Contact

Your questions and suggestions are always welcome!

Please do not hesitate to email us.

Stay connected

The newsletter

Sign up for the the Fraunhofer IIS Magazine newsletter and get the hotlist topics in one email.

Home page

Back to home page of Fraunhofer IIS Magazine