17/03/2022
Quantum computers become reality
Large sums of money are currently flowing into the development of knowledge and infrastructure for quantum technologies. To name two examples: two billion euros in the Federal Ministry of Economics and Technology's economic stimulus and future package adopted in 2021, and 15 billion dollars spread over the next five years in a current programme of the Chinese government. Commercial enterprises are also turning to the topic - especially since quantum computing capacities are becoming more readily available and new records in performance are constantly being set. And finally, large cloud providers such as AWS (with Amazon Braket) and Google already offer each of their customers the possibility to programme and use real quantum computers themselves.
In the following, our employee Eva Ess (Project Manager in the Automotive & Manufacturing sector) asks Thomas Klemm, our expert in quantum computing, exciting questions about how he assesses the potential of this new type of computing technology. Thomas Klemm is committed to the topic in the msg Group and advises companies on the possible uses of quantum computing.
Quantum computing is on everyone's lips as a "game changer" due to its expected future and value creation potential in business and politics. How are companies like BMW and VW currently positioning themselves on the market?
Our customers are also investing to benefit from the expected gigantic value creation potential. Volkswagen already started to deal with this intensively a good five years ago, on the one hand in selected specialist departments, and on the other hand by founding a specialised subsidiary, the VW Data Lab. BMW took a similar path last year and founded the QUTAC consortium with nine partner companies. And of course, the topic has also arrived in BMW's innovation management. This is shown by the most recent Crowd Innovation Challenge, which dealt with possible applications of quantum computing and in which we were involved. So things are happening.
Is this simply the latest hype? What is so special about this completely "different" kind of machine computing?
No, by no means is it "just the latest hype". It is:
- a different kind of machine computing, as you rightly say. Quantum computing is not just a bit faster, but fundamentally new and different in that it does not calculate with simple numbers - with bits and bytes - but with quantum physical states. The special properties of the "qubits", which serve as carriers of quantum information, open up fundamentally new approaches to familiar problems that are difficult or impossible to solve today, with completely new types of algorithms, for example.
- the main difference to today's computers lies precisely in these quantum algorithms. Thanks to the specific properties of the qubits, a quantum algorithm can process all possible solutions to a problem simultaneously. This leads to an enormously high computational parallelism. Example: If I want to find a certain telephone number in a telephone book with four million entries, then in a classical algorithm I have to read an average of two million telephone numbers and compare them with the number I am looking for. With a quantum algorithm (the well-known Grover algorithm), a good 2,000 steps could suffice.
- This gives rise to completely new hopes associated with quantum computing. We are thinking about problems that require extremely high resource input for today's computers due to their computer architecture or are even unsolvable within a reasonable period of time. The hopes are well-founded to crack such problems with quantum computing and its fundamentally different way of computing - even if we still face great challenges on the way to their fulfilment. Looking at the current progress, I am convinced that we will already see a number of groundbreaking applications within the next two to five years. By the way, we are also talking about completely different applications here that go beyond machine problem solving, such as quantum teleportation or quantum cryptography, i.e. the tap-proof transmission of information.
Back to computing: How can one imagine the way a quantum computer calculates in order to increase today's computing power to such an extent?
To do this, we need to delve a little into the mysteries of quantum physics.
A quantum computer does not serially process a sequence of individual numbers like a computer today (like any kind of Turing machine), but works - as I said - with quantum states of microscopically small systems. As you surely remember from your physics lessons, there are things like Heisenberg's uncertainty principle, wave-particle duality and other scary concepts. The background to this is: As long as I do not ask a quantum system about its state by means of a measurement, it can assume as its state any mixture of all possible states. One can then "calculate" with this mixture by physically influencing it, for example by rotating the atoms used for the quantum computer (something like in a nuclear spin tomograph). Only after you have performed such operations do you measure and thus force the system to assume a certain state, a result. Sounds pretty weird, and when Richard Feynman - one of the great physicists of the 20th century - made the corresponding proposal for such extremely complex physical experiments almost exactly 40 years ago, one could only dream of it.
What is the concrete advantage of calculating with quantum states (instead of "real" numbers)?
The quantum computer works with such mixed quantum states. Mathematically, this means nothing other than that I can calculate with linear combinations of the "pure" states. In a single computing step, all possible solutions are then processed simultaneously. This is the basis for the enormous parallelism that a quantum computer is capable of.
A small numerical example: How complex the quantum states of a quantum computer can be depends on the number of qubits mentioned that it can work with. This is "simply" the equivalent of the bits of normal computers, because a qubit has two quantum states (0 and 1, so to speak). Now, with 1,024 of these qubits, I can consider possible values in a calculation at the same time; with 20 qubits, there are already more than a million possible values! Today we are at about 100 qubits, next year IBM wants to break the 1,000-qubit mark, and by the end of the decade we want to be at one million qubits. You can now calculate for yourself how many solutions can be calculated simultaneously.
Source: https://www.fokus.fraunhofer.de/de/sqc/quantencomputing
At any rate, this is the reason for the high parallelism of quantum computing, and why problems could now come into the realm of solvability that far overtax today's computing capacities - especially those notorious problems that the computer scientists among us have come to know as "NP-hard". But also a large part of the encryption methods used today in e-commerce, cryptocurrencies and electronic communication would be vulnerable (which is why many government agencies are interested in quantum technologies).
Key Facts by Stephan Melzer, Executive Project Manager, msg
- Quantum computing can revolutionise IT like no other technology because it allows for fundamentally new computing.
- Quantum computers will be ready for industrial use in the near future.
- If complex algorithms are used in my core processes, it makes sense to think about possible application scenarios today.
It is precisely the applications of quantum computing that we want to address in the continuation of the interview.