© Daniel Isbrecht, RAZ Verlag

25 March 2019

Data-based research is transforming Berlin into an engine for the development of manifold AI applications.

The image of an underground canal system and that of a clinical server room could not be more different. Indeed, apart from sharing a monotonous background noise, the two have very little in common. So it’s surprising that Professor Dr. Volker Markl drew precisely this comparison at a recent conference of international experts in an attempt to explain what it is that he does. Markl sees himself as a kind of plumber, describing his work in very vivid terms: “You need pipes, high-pressure pumps and treatment plants to be able to do great things with water, like building fountains and cooking delicious meals. It’s my job to make sure everything runs smoothly with regard to the water supply.”

Of course, the water in this analogy represents an entirely different raw material – one that is in high demand in our increasingly digitalized era, namely data! Markl is head of the Berlin Big Data Center (BBDC) and co-director of the Berlin Center for Machine Learning (BZML). These two Berlin institutions work closely together to examine the fundamentals of what is commonly described as artificial intelligence (AI). This term is not always the first choice in research circles, notes BZML spokesperson and BBDC co-director Prof. Dr. Klaus-Robert Müller, one of Markl’s colleagues and himself an international though leader in the field: “AI is a word designed to give the general public a vague notion of what we do,” he says. As Müller notes, the foundation of AI is, in fact, machine learning and database management, and ultimately also mathematics and computer science.

Thanks to the close cooperation between the BBDC and BZML, Berlin now holds a truly leading position in Germany and also plays a pioneering role far beyond its borders. “All machines learn from data. Period,” says Müller, emphasizing the universal potential of the work being done in the capital. The ultimate aim is to procure an adequate amount of this data resource, handle it responsibly, manage its myriad of individual pieces and thus “feed” machines in such a way that they can analyze it in an optimal manner and bring connections to light. “This doesn’t mean sitting in front of a computer all day in a quiet room,” notes Markl, correcting the pre-conceived idea many have of his job. “In fact, it is a very communicative profession. It’s varied and exciting and has plenty of room for ideas and visions. We launch projects, write scientific articles, assist in the emergence and growth of startups, cooperate with industry and generally get around a lot.”

Markl and Müller, two multiple award-winning scientists, have already celebrated a number of successes, as their teams regularly function the engine for a number of popular developments. For example, the “Apache Flink” application used worldwide today owes its existence to Markl’s research. This app links computers to one another and manages their cooperation in the analysis of giant amounts of data. Müller’s working group alone is responsible for the founding of 15 companies and the creation of 400 new jobs in the capital. Among other applications, his research has made its way into the software used in image and speech recognition, in autonomous driving for the auto industry and in the field of medicine, where it has assisted in the analysis and forecasting of cancer cell development and used in brain-computer interfaces designed to help locked-in patients make their wishes known. “Centers of expertise are platforms through which we exchange knowledge with different disciplines,” says Müller. “We are in constant dialogue with the business community and ultimately also with society. Indeed, we continually make an effort to help people understand what it is that we do exactly.”

As Müller notes, the ultimate goal is to generate enthusiasm here in Berlin and make the opportunities understandable to all. It is thus also important to transport these ideas to the political realm. Last summer, Chancellor Angela “Innovation cannot be planned” Prof. Dr. Klaus-Robert Müller 17 Merkel organized a gathering of AI experts to which Müller and Markl were also invited. The core of the meeting focused on developing a national strategy that would bundle the potential advantages of digitalization. This factor is a key priority for the two Berlin-based scientists, because their location will only be able to flourish if it receives support from both the federal and state levels of government: “We already have the expertise in Berlin, but it must be secured for the long-term in a sustainable way,” notes Markl. He argues that the most important thing is that groups of specialists have time and space for open-ended cooperation, alongside steady and simultaneous funding, an attractive research environment and opportunities to foster young companies. “Most people take wwhat they know and extrapolate that. But that’s not how innovation works! It cannot be planned,” underscores Müller. On the contrary, in the case of broad and ongoing work, sometimes the only thing you can hope for is something unexpected to happen; something that causes “the whole world to look different.” As Müller says, “This is basically our core business.”

At this point, it’s quite clear that there is tremendous big-data potential waiting to be discovered and canalized with the help of data-based research from Berlin. It’s also quite clear that this will involve a certain race with the competition. The United States and China have indeed taken the lead in the industry. However, in the past several years, Germany – and especially Berlin – has set out on a veritable sprint. “I believe we’re doing pretty well,” says Markl modestly with regard to his own performance. Like Müller, he also emphasizes the growing climate of innovation with new companies and more capital.

Markl also points out that the infrastructure for founding companies has grown significantly. This is a sector that ends up re-inspiring itself over time, as first-generation role models, mentors and business angels pass on their knowledge and industry contacts to the next group of young innovators. Markl also emphasizes the work being done by Berlin’s universities to educate students in the field of big data, which provides students not only with a scientific basis but also with the necessary business acumen. In turn, this is great for recruiting: “There are three times as many fantastic people out there than jobs,” reports Müller. This is a true luxury considering the high market demand. Still, both funding and personnel resources are ultimately finite, he admits, a little tongue in cheek: “I myself want to be able to slow down at some point, become a wise old professor and tend to my roses.” Although the image of this “teacher of machines” in his rose garden is meant, again, more as a symbol, it nevertheless corresponds to the analogy mentioned above. Indeed, roses also need water to grow and unfold their full splendor.