Quantum Computing

Both conventional computers and quantum computers perform functions with bits, but they differ significantly in other ways. A bit in a standard computer can only assume two states, namely “0” and “1”. However, a quantum bit or qubit of a quantum computer assumes both states simultaneously in a superposition. This ensures that the computing capacity of a quantum computer is much higher than that of a conventional computer because data can be processed simultaneously. And with each additional qubit, the computing power grows exponentially.

Superposition, which is the overlapping of states, also leads to qubit interference. This means that the states reinforce or weaken each other. Another concept that allows quantum computing to deliver high performance is the entanglement of several qubits. As a result, the states are no longer independent, but are coupled together – if one qubit changes, so does the other.

These three features of quantum computing (superposition, interference and entanglement) ensure that quantum computers can process data more efficiently. This enables them to work with complicated datasets and, in the future, means that they will be able to perform previously unsolvable computational tasks.

Unlike conventional computers, the processes in quantum computers do not take place at the electronic level, but instead at the physical particle level. Special frameworks are necessary so that the quantum mechanical properties of a quantum computer can be maintained. For example, quantum computers based on superconducting qubits consist of several chambers, the lowest of which is cooled to a temperature close to absolute zero (-273 °C). Even minor temperature changes or radiation could influence the qubits. As yet, it has only been possible to actually keep a qubit in the state of superposition for a short time – but maintaining this state is a prerequisite for quantum computing to function properly. The number of qubits is also currently very limited. In a few years, there will probably be quantum computers with thousands of qubits, but this still may not be enough for complex calculations.

This means that a lot of fundamental research is still necessary at this stage so that quantum technologies can be used reliably in the future. However, interest in quantum computing is growing, given the potential for quantum technology to make groundbreaking developments in many industries due to its phenomenal computing power.

Currently, quantum computing is still in its infancy. Many companies and institutions are researching algorithms that could already benefit from today’s young quantum computing devices. Many scientists are also working to improve the underlying hardware and develop powerful simulation environments for quantum computers.

Additionally, the Fraunhofer-Gesellschaft is conducting research on quantum technology together with the cross-institutional Fraunhofer Competence Network Quantum Computing, and as a result is making significant progress in quantum technology’s development in Germany. To carry out this research, since January 2021 the Fraunhofer-Gesellschaft has had access to the IBM Q System One quantum computer, which is operated by IBM in Baden-Württemberg. It has 27 superconducting qubits and can also be used by external partners, such as companies and universities.

Focus on software for quantum computing

At the same time, research and industry are increasingly beginning to look at software for quantum computers. The Fraunhofer Institute for Cognitive Systems IKS is focusing on this area and is part of the Bavarian Competence Center for Quantum Security and Data Science (BayQS) and the Munich Quantum Valley (MQV).

The aim of the Competence Center (BayQS) is to address the three aspects of cybersecurity, reliability and optimization at this early stage in the development of quantum computers and ensure that they are given due consideration from the ground up. Its work focuses primarily on software issues related to quantum computing. By interconnecting the three topics under the umbrella of the BayQS center, it is laying extensive foundations for maintaining quality in quantum computing, something that will help make this technology widely applicable. In addition to Fraunhofer IKS, the Fraunhofer Institutes AISEC and IIS are also involved.

The Munich Quantum Valley is working on quantum computing demonstrators based on three different technologies: neutral atoms, superconducting qubits and ion traps. The entire software stack is also being co-developed through to the end of the application, a process that Fraunhofer IKS is involved in. In this case, the focus is primarily on reliable quantum computing algorithms for industrial, pharmaceutical and chemical applications.

As part of the BayQS center, Fraunhofer IKS is conducting research on safe software applications for quantum computers. Since a quantum computer is only really useful if you can rely on its calculations, quantum computing needs not only sheer computing power, but also robustness in order to be successful.

Fraunhofer IKS is contributing its expertise in the field of safety to two BayQS projects that combine quantum computing and artificial intelligence (AI):

Reliable quantum computing-based AI for medical diagnostic tasks

In the future, quantum computers are expected to improve diagnoses based on image data, such as MRI or CT scans. Traditional AI methods require a significant amount of data in order to reach reliable conclusions. Quantum computers, on the other hand, could provide a more reliable result with the same amount of image data. Fraunhofer IKS is conducting research into this to improve the diagnosis of brain tumors, for example.

• Quantum based AI in medicine

Quantum computing-based certification of neural networks

Neural networks quickly become complicated when used in real-life applications, making them difficult to certify. As part of this project, Fraunhofer IKS is investigating how quantum computing can be used to make autonomous, networked systems safer. The primary focus is on providing verifiable guarantees for certain properties of complex neural networks.

• Quantum computing-based certification of neural networks