Quantum Computing coming soon
Professor Sabina Jeschke is one of Germany’s leading experts in quantum computing and plans to make these supercomputers fit for everyday use with her startup company called Quantagonia. As a physicist and mathematician, she teaches Information Sciences in Mechanical Engineering at RWTH Aachen and Entrepreneurship at TU Berlin. As a member of the executive board of Deutsche Bahn AG, she was responsible for digitalization and technology from 2017 to 2021.
You’re planning to bring quantum computing into everyday life. Today’s quantum computers are big, reputed to be error-prone and energy-hungry and will only start operating at temperatures close to absolute zero. Those aren’t exactly perfect prerequisites for a technology that’s supposed to be suitable for everyday use, are they?
It’s true that quantum computers are still error-prone at the moment but they’re not energy-hungry. Current quantum computing systems have to be run at millikelvin temperatures, in other words at minus 273 degrees centigrade (–459.4 degrees Fahrenheit) but because the actual computing effort is minimal, because many people can share these computers in a cloud and because the computing speed of this technology is so much higher its energy balance is often already much better today than that of traditional computers. Plus, there are some prototypes now that run at room temperature and are less error-prone. We anticipate marketable systems to be available by 2028 or 2029. Consequently, while today’s data centers are consuming increasingly larger amounts of energy, quantum computing can turn the trend and help us in areas like sustainability. That’s why it’s possible that in the future everything that’s algorithmically feasible will be processed on quantum computers and classic computers will be used primarily in areas where algorithms cannot – or only inefficiently – be transferred to quantum computers. By the way, that also provides new impetus to so-called edge artificial intelligence, in other words the use of artificial intelligence (AI) at the edges of networks. The mobile sensors that are frequently used there often have just tiny batteries and therefore insufficient energy for complex computing tasks. However, for quantum computing, the minimal amount of energy that’s required can be obtained as a waste product from the vibrations of machines. That enables all-new AI applications.
"One of the major trends in quantum computing going forward is moving away from the low temperature and error-proneness."Professor Sabina Jeschke
In classic computer technology, decades passed between the mammoth ENIAC vacuum-tube computer and chips using conductive traces with nanometer widths in every cell phone. Are we going to see similar kind of progress with quantum computing? Will the iPhone become a Qphone?
Yes, you can clearly put it that way. One of the major trends in quantum computing going forward is moving away from the low temperature and error-proneness toward more robust systems – ideally operating at room temperature. Once error-proneness has ceased and I no longer have to cool quantum computers in large underground centers, I expect to see devices that I can carry around with me, such as a Qphone.
Every new technology affects existing ones. Will quantum computing become a disruptive technology and will the big data centers like those operated by Google and Amazon soon no longer be needed?
Quite the opposite is the case: We’re going to see that the big providers of high-performance computing will be buying quantum computers, putting them in their data centers and offering that computing power to their customers. Schaeffler, for instance, uses Microsoft’s Azure cloud computing platform. In that cloud, you can already access a diverse offering of quantum computing today. Amazon does that in very similar ways with AWS, so quantum computing will in fact become a disruptive technology because you can use it to address problems, such as complex simulation at high speed, that you couldn’t do before. The big data centers are not going to be the losers but will put quantum computers next to their existing computers. We’re going to see a very heterogeneous computer landscape in the data centers and we’re going to distribute computing tasks precisely to the computing architecture that’s ideally suited for them.
Whereas current digital computers only compute with the values of “0” and “1”, quantum computers achieve what seems to be a paradox: In addition to “0” and “1,” their informational units, the qubits, can assume any intermediate states – and even simultaneously. The computers take advantage of the laws of Erwin Schrödinger’s quantum mechanics using ions or electrons as qubits for instance. When interleaving several qubits into a larger unit and asking each qubit a subquestion of a complex overall task the quantum computer will find the best possible combination of answers by taking all subquestions into account. As a result, a quantum computer in a search for new materials or medications can compare millions of molecular combinations within fractions of a second. At this juncture, quantum computers only have few qubits and are prone to computing errors and interferences, but the development is progressing rapidly.
One of your research objectives is to harmonize quantum technology with conventional information sciences and existing programming languages. How do you teach something like that to existing software architectures that think in zeros and ones and don’t tolerate any intermediate states?
In fact, the exciting question here is: What am I going to do with existing codes? In science and in business, we’ve been using some codes for 40 years and keep developing them because they’re stable and work incredibly well. How can I manage to preserve our grown structure of simulation, optimization methods and artificial intelligence? In our startup Quantagonia, we’re building a “Low Level Virtual Machine” or LLVM for short, a virtual computer that exists on another system only as software. We then feed that with existing code. Simply put: “input X86 code, output Q code,” so we transfer the existing code to a machine-independent intermediate level and distribute it from there to the various computers with various modifications. In that way, we create a “middleware” that makes the existing software compatible with quantum computing and simultaneously enables an ideal utilization of heterogeneous resources in a data center by means of dynamic allocations.
Your research areas also include transportation and mobility as well as the Internet of Things (IoT). Will the autonomous automobile of the future with a quantum computer on board find a better route through big city traffic? What can a quantum computer do that conventional circuit logic in navigation systems couldn’t do?
Fleet intelligence. At this juncture, you can use existing computers to optimize individual vehicles or small groups but not the entire traffic of a big city. The more quantum intelligence is in the navigation systems the better their predictions become, plus, the better the drivers’ compliance with the systems’ predictions. That, in turn, decongests traffic. Traffic jams, not least, develop because, as a human, I don’t have enough possibilities to see how my behavior contributes to congestion in the first place. In the future, thanks to quantum computers, I’ll know that I’d better leave ten minutes later, that will reduce overall congestion and that I’ll arrive earlier in spite of having left later. The prospect is for autonomous vehicles, which will become increasingly prevalent in the coming years, to be fully controlled by fleet intelligence.
What can corporations like Schaeffler as one of the globally leading automotive and industrial suppliers expect of the quantum revolution? Are you seeing new business segments and sales markets in mobility? Or will the revolution tend to take place behind the scenes, for instance in research and development using quantum computing or in all-new manufacturing technologies in Industry 4.0 factories?
Quantum computing, for instance, will provide new opportunities for material research, tribology and the search for new lubricants. Going forward, it will be possible to create simulation environments in which parameters can be changed in real time and the results are available immediately. That hasn’t been working so far because such calculations on the molecular level have response times ranging from days to weeks. In the future, quantum computing will do that in real time. I’ve already addressed edge artificial intelligence, where unique selling propositions in the automotive sector may emerge for companies like Schaeffler, for instance with new components using artificial intelligence even without a cloud connection via 5G mobile telecommunications. Such systems could then assume complex control tasks in vehicles offline in real time thanks to quantum intelligence.
How do you judge the industry’s innovative prowess in terms of quantum computing? Are the opportunities of this technology adequately recognized or would you occasionally like to see a faster pace here?
Germany is a country of engineers and technologists. Even so, we’ve missed incredibly many trends in the area of technology, with cell phones and, very generally speaking, with hardware. We’ve also recognized the potential of AI far too late. But I do see a tendency to embrace change. For instance, the EU and the German federal government have launched major funding programs for quantum research. At Schaeffler, I’m currently engaged in many conversations about quantum computing on the factory floor, quantum computing in the product and in relation to talent development. That’s another reason why my outlook is generally positive.