At the moment everyone is talking about quantum computing – it is said to be the quantum leap in IT. Recently, Forschungszentrum Jülich (FZ Jülich) and Google announced that they would collaborate on research into quantum computers – as we reported. All the more reason for us to ask more detailed questions.
Prof. Dr. Kristel Michielsen, Group leader “Quantum Information Processing” & Member of the Division “Computational Science”, FZ Jülich,
Prof. Dr. Dr. Thomas Lippert, Director of the Institute for Advanced Simulation & Head of Jülich Supercomputing Centre, FZ Jülich &
Dr. Markus Hoffmann, Global Quantum Computing Practice Lead, Google Cloud
in an interview with Innovation Origins.
Could you please briefly explain the difference between supercomputers and quantum computers to our readers?
Prof. Dr. Dr. Thomas Lippert: Today’s supercomputers, like for example our JUWELS at the Forschungszentrum Jülich, consist of many computer hubs that are networked in such a way that they can work together on one problem. They are able to generate an enormous amount of output from this parallelism. JUWELS, for instance, could take on up to 60,000 laptops if they were properly connected to each other. (…)
Quantum computers generate their huge output from a completely different kind of parallelism: (…)
While the output of conventional supercomputers is rising substantially in proportion to the number of CPUs connected, people hope that if you add an extra qubit, the output of the quantum computer will double in size.
“There is talk of an exponential increase in performance. This is precisely what makes the quantum computer so special.”
A nice example concerns the route planning for a commercial agent
[Notes: as in those who want to calculate the shortest route when going to several locations]: If you add another destination, the calculation effort for finding the optimal route increases by a factor of around two.
This example can easily be applied to the realm of traffic optimization. Such as optimizing the route planning of 10,000 taxis in Beijing, which apparently is a major problem, one that Volkswagen is addressing with quantum computers. Or think of the optimization of air traffic routes. We believe that we will soon see these technologies in our everyday lives as well. Think about the flow of commodities or the optimal positioning of products in the supermarket – optimization is needed everywhere.
In the context of quantum computers the terms qubits/quantum bits are used – how do they differ from bits?
Prof. Dr. Kristel Michielsen: Bits can have two values, e.g. 0 and 1. An element of a conventional memory is e.g. one such bit. These bits are processed using transistors.
Similarly, qubits are the elementary memory elements in quantum computers. Qubits may also be manipulated depending on their technological design, e.g. via microwaves. With the help of the interaction between the qubits, so-called Quantum logic gate [specialized quantum calculation systems] are constructed – just as conventional gates are formed in transistors.
In conventional computers, these gates are logical conduit switches that process information on the quantum computer which we refer to as the quantum state, wherein we hope they do so according to the laws of quantum mechanics.
The sum of all these programmable steps is called an algorithm or a quantum algorithm.
While conventional bits are solely 0 and 1, quantum bits, which are often represented by small arrows, are able to adopt all conceivable positions of an arrow [Notes: the tip of an arrow can be aimed at any point on a single sphere]. Configuration of the positions of several Qubits then corresponds to the superimposed states as referenced above. Except when one measures these, the individual Qubits revert back to the values 0 and 1, which corresponds to the set arrow’s position in relation to the sphere’s south pole or north pole. The result of this computation process can then be read from these arrow positions.
[Notes: As Qubits adopt a certain condition/location only when you want to measure them. Imagine a billiard ball lying on a table, which will be measured by another billiard ball. As soon as the rolling ball hits the ball to be measured, the original ball changes its position on impact… However, physicists at FZ Jülich are very critical of these types of simplistic comparisons]:
“Often one tries to explain these overlays and processes with conventional analogies. We think that this is arguably more confusing than revealing, because if a quantum state or a quantum computation could be represented in such a way, one would simply be able to do it with conventional computers.”
There are a few companies who are also researching quantum computers – why did you choose Google to work with you?
Prof. Dr. Dr. Thomas Lippert: In the field of quantum computing, Jülich is by no means bound to a sole partner. Google is the first of hopefully several other partners. JUNIQ, the Jülich Unified Infrastructure for Quantum Computing, plans to collaborate with all major companies in this area. One example is the Canadian company D-Wave Systems. Others are planned. Very important for us is the teamwork within the European Quantum Flagship Project, where we are involved in the OpenSuperQ subsidiary project and want to take control over the operation of the European prototype.
The deal with Google is a partnership that promotes the development and deployment of quantum computers. Google is a leader in the construction of gate level quantum computers [Notes: In 2018 Google introduced the world’s fastest processor with 72 Qubits]. We, together with our partners and users within the quantum computing field, will gain access to Google’s quantum machinery and will be supported on issues concerning software applications and algorithm research.
Mr. Hoffmann, there are many universities doing research in the field of quantum computing – what was the reason for Google choosing Jülich and not working with a university in the USA?
Dr. Markus Hoffmann: The Google Quantum AI Group maintains multiple collaborations and partnerships with academic institutions and research laboratories worldwide. An example of this is the UK-based Prosperity Partnership with the University of Bristol and the UCL Quantum Science and Technology Institute in London. We have had a long-standing cooperation with Jülich researchers over many years, so formalizing a partnership was a logical step.
(To everyone): Isn’t the German-American cooperation going up against the expensive EU research project Quantum Flagship, which was launched last year and aims to strengthen the European market in this field?
Prof. Dr. Kristel Michielsen: Not at all. Jülich is a partner in the European Quantum Flagship Project, where we are involved in the OpenSuperQ subsidiary project and intend to take over operation of the European prototype in due course. We get great support from Google when it comes to software applications and research on algorithms for gate level quantum computers. This helps us a lot in the OpenSuperQ project. Incidentally, there is currently far more intellectual property flowing from the USA to Europe than vice versa.
The location of the first European quantum computer built by the research project will – as you have just pointed out – be Jülich. When is this expected to be completed?
Prof. Dr. Dr. Thomas Lippert: OpenSuperQ, the subsidiary project of the EU Quantum Flagship Project, which is developing the quantum computer with superconducting qubits, would like to build a prototype with 50 to 100 qubits by 2021 [Notes: for comparison: 45 qubits roughly correspond to the memory size of the largest current, conventional supercomputer and 50 qubits represent, or so it is presumed, the superiority in computations]. This will then be integrated into JUNIQ and made accessible to European users.
Whether using superconductors, laser-powered ions or photons – there are several technologies that are preferred by various manufacturers. Google, with its Bristlecane, stands for superconductors. To what extent will you also be researching alternative technologies?
Prof. Dr. Kristel Michielsen: In Jülich, research is carried out on superconductive, semiconductive and topological qubits. Other technologies are under discussion and a decision will be made in the next few months.
Could you please name some of the areas of application for the quantum computers that have been studied in Jülich – and in particular those areas that have been explored in cooperation with Google?
Prof. Dr. Kristel Michielsen: Areas of application are found predominantly where optimization plays an important role, and in particular where optimization problems are so complex that supercomputers cannot get any further ahead. And these are many areas, for example in the design of pharmaceuticals, in production planning, in commerce, traffic, aviation, etc.
Quantum computers operating together with supercomputers in hybrid mode are likely to achieve far better computing and energy efficiency than supercomputers on their own. Another important field of application is the simulation of quantum systems. It was here that the founders of the technology, like Richard Feynman, foresaw the most promising applications.
How do you see your goals? When should the first quantum computer become marketable, or what are your intermediate goals?
Prof. Dr. Kristel Michielsen: Predicting the future is always very difficult. Surely we will not all of a sudden be able to operate our own quantum computer at home. Quantum computers are complex entities and will be located at a data center, where they will be operated together with supercomputers in hybrid mode for optimal efficiency. This is our goal at the Jülich Supercomputer Center.
Some quantum computers are already on the market. For example, IBM already offers customers remote access to gate-level quantum computers, and D-Wave Systems is selling its quantum annealer (NB: quantum computer for calculating optimization problems].
A quantum computer in the wrong hands could endanger IT security – to what extent do you include this issue in your research?
Prof. Dr. Dr. Thomas Lippert: In that case, then, it is particularly important that we at Jülich are familiar with these types of systems and that we study them in detail. You always have to be one step ahead of the bad guys. We think we have quantum computers in the right hands. And IT security is of course always of paramount importance to us, especially when it comes to our supercomputer. Cyber security is an important area of research for several of our partner institutes within the Helmholtz Association
What do you now recommend various companies do – should they be concerned with quantum computers and new cryptic processes and perhaps prepare themselves for the next ‘quantum leap’ in IT?
Prof. Dr. Kristel Michielsen: Yes, companies need to prepare as quickly as possible for the deployment of quantum computers as a means of solving specific numerical problems. The competition never sleeps. Quantum Computing is attracting increasing interest from industry and science groups that use high-performance computing (HPC) for their applications. These pilot users of quantum computing are primarily interested in testing whether the available quantum computer technologies are suitable today or in the foreseeable future for solving the problems relevant to them. At JUNIQ, our goal is to support these industrial users. Another important goal is to train doctoral students and company employees in the field of quantum computing.
And last but not least: What is your vision? To what extent will quantum computers revolutionize our personal computing lives?
Prof. Dr. Dr. Thomas Lippert: As already mentioned, a potential first practical application could be the traffic optimization of entire cities for autonomous driving, as is currently successfully being researched by Volkswagen. Further examples are the dynamic optimization of flight plans or the analysis of satellite data, a domain in which the German Aerospace Center is active. At Jülich, we are convinced that quantum computing is one of the most promising future technologies in the fields of computer simulation, artificial intelligence and industrial optimization. We are working hard on the integration of this technology into our modular supercomputer environment.
If the subject of quantum computing has piqued your interest, you can find a good overview on quantum computers here or here.
A selection of our articles on quantum physics can be found here.