When Frank Leymann first heard about quantum computing, he was skeptical. Then he studied the subject intensively and became an expert and advocate. The Professor of Computer Science at the University of Stuttgart explains why quantum computing is on the threshold of being used for new business opportunities - and why companies should deal with it today to benefit from it tomorrow.
Frank, what is the most fascinating aspect of quantum computing for you?
It’s the ability to solve problems that the community thought could never be solved, in a reasonable amount of time. For example, prime factorization. We learned it at school. If you have a number and you want to determine the prime factors, it sounds easy, but it is untractable. You can't solve it practically with classical algorithms. An error corrected quantum computer can do this within a couple of minutes. And it adds another dimension: Precision. Quantum computers can solve problems with much higher precision than is traditionally possible.
Why is it so powerful?
Quantum computers are not dealing with bits with only two different values, zero or one, but with quantum bits - or so-called qubits. A qubit consists of any combination of zero and one, which is called superposition. It's an infinite number of potential values at the same time. And when you manipulate them, you are manipulating this infinite amount of data. Also, there's a very subtle phenomenon that even Einstein could not really accept. It’s called entanglement. It means that a qubit can be brought into a very close relation to other qubits. This means if you are manipulating one qubit, the other qubits are also manipulated at the same step.
Some people suspect it’s just a mysterious hype. What do you reply?
I was skeptical, too, at first. But as I studied the subject and dug deeper, I understood the relevance and I became excited. I began to understand that this was a real thing. There are still serious hardware issues to overcome, no doubt about it, but the progress is much faster than we ever expected.
What will be achievable in the next couple of years, so that business users can really make sense of this powerful tool?
The first achievement will be the simulation of physical and chemical systems, which is important for material science. Another one will be in the realm of medicine, where you try to find new molecules that can heal diseases with targeted medication. The third realm will be - and is - machine learning. It is something that you can do quite well on quantum computers at a reasonable scale. Fourth, the so-called portfolio optimization problems are already being solved in reasonable sizes today on existing quantum computers.
What is the concrete scenario in material science?
It's not a secret that automotive companies are working on batteries for electric vehicles. Today’s batteries are limited by a lack of natural resources, such as cobalt. There won’t be enough of it for all the electric vehicles and their batteries that will be built in the next years.
That means that material scientists are trying to replace cobalt by other substances?
Exactly, and for this purpose, you need to use a simulation. Simulation takes a very long time, even on high-performance computers. First steps already show that this kind of material research can be done very quickly on quantum computers.
Do you see any applications in the realm of the Industrial Internet of Things?
Of course. Imagine you have zillions of sensors out there and they are spitting out tons of data. Speed is important to pre-process and process this kind of data. That’s where quantum computers come in, because not only do they have better precision, but they are also much faster. Furthermore, you can establish uncrackable protocols between the sensor and the processor. Cryptography as well as speed are where the IoT will really benefit from quantum computing.
When will quantum computing applications be ready for business practice?
It depends. Optimization problems can be solved to a reasonable extent in the next three to five years, I guess. The next problems would be from machine learning. We have to wait at least another five years before we can solve huge problems. And then security: It is especially important because quantum computers can be used to crack today’s keys. Basically, we need to apply quantum technology to secure our communication channels, which requires that we must change our entire communication infrastructure. And this will take at least ten years.
How should companies position themselves today to be ready for these breakthroughs?
They don't need to understand quantum mechanics, but they do need to understand the completely new paradigm to see the potential for their particular business area. They also need to rethink what it means to work with a quantum computer.
If you are a traditional computer scientist, you are used to testing your programs. But quantum data cannot be copied, i.e. stuff like debugging is a problem. Also, you need to interact with quantum physicists who invent new quantum algorithms. You need to touch all your development processes to get ready for quantum computing.
My recommendation: Start using quantum computers now. Start with a few people who are interested in advanced technology. There are quantum computers available for free or a little money in the cloud from Amazon Web Services and IBM. Take one or two key problem areas in your business and try to implement a solution based on quantum computing so that you can gain experience. And when quantum computers are ready to be used in practice in a few years, you’ll have the team that can basically work with them. Very likely, you will then have an advantage over your competitors.
But to get this straight: Quantum computer will not substitute our current IT?
Even a very famous quantum algorithm like the so called Shor-algorithm, which can be used to crack keys, is a hybrid algorithm. It requires classical preprocessing steps. Then you use the quantum computer a little bit, and then post-processing is required. A quantum computer is great for certain things that are very hard to do in a classical way. You need to retrieve data from a hard disk. You need to massage the data to prepare it to be manipulated in the quantum computer. Then you go to a quantum computer to do a single step to analyze the data, and then you find out that you need to go back to a classical computer.
So, you are saying, using quantum computers to build new applications is inherently an integration challenge?
That’s right. Using a quantum computer, you have to integrate a lot of classical processing with advanced specialized processing. Quantum computing is hybrid. It will be a big integration challenge, and companies should be prepared for that. Integration technology is of paramount importance. It can't be overstated. Again, my recommendation for companies is: start now! Quantum computing is opening up a new age of technology.
Frank Leymann is computer scientist and mathematician. After his PhD in mathematics, he went to IBM Research and Development contributing to software products like DB2, Websphere or MQSeries. In 2004, he was appointed full professor of computer science at the University of Stuttgart where he founded the Institute of Architecture of Application Systems. He holds many granted patents in the area of software. In 2020 he became Member of the Expert Council for Quantum Computing of the German Government.