“I find it fascinating how quantum systems store information and perform calculations,” says Greplová. “During my postdoc at ETH Zurich, I noticed that quantum systems present some inherent challenges. For example, they cannot be simulated with a classical computer – which is precisely why we build them. But how can we understand and control these quantum systems and verify , what they do?”
Looking into the quantum computer with artificial intelligence
With his research group at TU Delft, Greplová adds AI to the application of quantum computers: “It is very complicated to simulate quantum computers, partly because it is difficult to model noise on the quantum scale,” says Greplová. “Traditional calculation models already exist for this. With AI’s enormous power of generalization, we can extend it to better describe and control the quantum hardware.”
Even looking inside a quantum computer could have disastrous consequences for quantum computing. Greplová realized that AI could provide a solution to this problem, which is a paradox analogous to Schrödinger’s cat. “In the original thought experiment, a cat is locked in a box together with a radioactive isotope that may or may not have killed the cat,” Greplová explains. “If we open the box and look inside, the quantum state collapses, and we thereby affect the quantum system. I realized that by ‘seeing’ the output of a quantum computer, we at least get a better understanding: AI can translate the complex data into useful information while interfering less in the quantum calculations.
Crowdsource sustainability ideas
Greplová had one of her first breakthroughs at ETH Zurich. “Our research group did an experiment on a small quantum computer using artificial intelligence,” she says. “Usually we need a whole team of experts to interpret complex data for us. Now we could completely outsource this time-consuming step to AI: our AI algorithm helped us control the quantum system, made it do what we wanted , and confirmed the results for us.”
What are Greplová’s ambitions? “I want to apply quantum computing to sustainability issues. What if we could come up with an alternative to plastic that degrades naturally – by putting all the chemists’ different ideas on a quantum cluster and letting them decide the best solution? That’s the big goal I want to contribute to with AI. If we can crowdsource these ideas around the world, we can solve such sustainability problems much faster.”
Scaling up quantum computing with AI
“This is what my research is about: I want to place quantum computers in the cloud using AI automation and gain global access to users. With AI we can control, read and interpret these quantum computers fully automatically. In addition, AI solutions, unlike standard programming use the data to determine which models are effective. This makes controlling quantum computers much easier to scale up.”
Before moving on to the actual application, researchers must first develop AI techniques specifically for quantum computers, a very complex challenge. Greplová: “This requires a quantum system consisting of millions of qubits – the building blocks of quantum computers that can simultaneously assume the values zero and one and any value in between. To put this into perspective: the current state-of-the-art systems consists of at most 50 qubits. I estimate that we will not be able to apply artificial intelligence to quantum computers on a large scale before the late 2020s.”
AI and physics: opposites attract
Another challenge is to convince the scientific community of the usefulness of AI in quantum computing. Greplová: “Artificial intelligence and physics are two completely opposite methods that I bring together in my research. In physics, you reduce a complicated problem to a simple model, while artificial intelligence tries to extract useful information from large amounts of data. Some physicists do not believe that AI will be able to interpret their data correctly. Especially at the beginning of my AI research, it was a challenge to convince physicists of the usefulness of my approach, but it is starting to dawn on them now.”
At TU Delft, Greplová works together with Ronald Hanson’s group, which within QuTech (a collaboration between TNO and TU Delft, ed.) is researching quantum internet applications. She works with Lieven Vandersypen and his team (also QuTech) on the Dutch Quantum Inpsire, the first European quantum computer in the cloud. “This unique quantum ecosystem was one of the reasons why I came to the Netherlands,” says Greplová. In addition to her scientific research, she is also interested in the control and monitoring of quantum technologies. To this end, she participates in the Global Future Council on Quantum Computing in the World Economic Forum. In 2019, she also co-founded the Virtual Science Forum, a platform that aims to make scientific events accessible to everyone.