Correlated errors in quantum computers emphasize need for design changes
Posted on
Quantum computers could outperform classical computers at many tasks, but only if the errors that are an inevitable part of computational tasks are isolated rather than widespread events.
Now, researchers at the University of Wisconsin–Madison have found evidence that errors are correlated across an entire superconducting quantum computing chip — highlighting a problem that must be acknowledged and addressed in the quest for fault-tolerant quantum computers.
The researchers report their findings in a study published June 16 in the journal Nature, Importantly, their work also points to mitigation strategies.
“I think people have been approaching the problem of error correction in an overly optimistic way, blindly making the assumption that errors are not correlated,” says UW–Madison physics Professor Robert McDermott, member of the Wisconsin Quantum Institute and senior author of the study. “Our experiments show absolutely that errors are correlated, but as we identify problems and develop a deep physical understanding, we’re going to find ways to work around them.”
The three-year, $4 million funding will allow the researchers to apply emerging tools to identify new materials and fabrication methods that can improve the performance of these systems.
“All of physics is quantum on some level, and quantum systems let you understand how physics works when you get to the cleanest, smallest, most isolated systems,” says Shimon Kolkowitz, assistant professor of physics at the University of Wisconsin–Madison and lead investigator of the grant. “We think that quantum computing, and quantum technologies more generally, are a really promising area of technological development and research.”
Quantum systems — which make use of single atoms or electrons and the quantum mechanical properties that govern them — have the potential to push boundaries in such areas as computing, precision sensing, and secure communications.
Quantum computers, for example, allow scientists to simulate quantum mechanics in ways that classical computers cannot. But, the computing power of quantum computers has not yet exceeded classical ones.
A scanning electron microscope image of a gate-defined double quantum dot qubit fabricated by the Eriksson group.
A limiting factor in quantum computing power is the number of qubits, or quantum bits, that can be strung together. Like bits in a classic computer, the more qubits in a quantum computer, the more the computing power. And the limiting factor in how many qubits can be connected with each other while remaining in the fragile quantum states required to perform a computation — called “coherence” in quantum lingo — is their resistance to external environmental factors, or “noise” that may cause them to “decohere.”
However, researchers have found that the materials used to make the qubits themselves generate a lot of this noise.
“People for quite a while have seen this noise, treated it as a fact of nature, and tried to design around it. But no one really knows what it is or how to get rid of it,” Kolkowitz says. “Even more fundamentally than just understanding or reducing this noise, we think that if you can reduce or ultimately eliminate this noise, it actually opens up the design space for the kinds of qubits you can build, and that will make it much easier to wire qubits together.”
With the DOE funding, Kolkowitz, along with colleagues at WQI and the Livermore National Laboratory, seeks to first identify the nature of the noise and how specific materials contribute to it, and then to develop ways to reduce it.
Atom-size quantum defects in a diamond, imaged using a confocal microscope in the Kolkowitz lab.
Work in Kolkowitz’s group, as well as that of Victor Brar, assistant professor of physics and co-investigator on the grant, has led to the development of quantum sensors that allow the researchers to characterize things like magnetic fields at the nanometer scale, or to see how single atoms are arranged in various materials. Part of the DOE funding will be used to continue improving these sensors.
Kolkowitz and Brar then want to use their sensors to identify the noise affecting qubits designed by UW quantum computing researchers Mark Eriksson and Robert McDermott.
“And then we can work in a feedback loop, where, for example, Robert McDermott makes samples and characterizes their performance, then we study the noise limiting that performance with these quantum and nanoscale probes to figure out what’s happening on the microscopic scale,” Kolkowitz says. “Then, we give that information to our theory collaborators here and at Livermore who build models and simulations based on what we’ve measured. And then Robert can use what we’ve learned to design and make new samples to see if we’ve improved on these issues.”
An image taken with a scanning tunneling microscope of a single atomic defect in graphene
Trying to identify sources of this noise is nothing new, but what Kolkowitz finds most promising about the work funded through this grant is the development and application of new sensing technologies.
“These emerging tools that use quantum states and quantum systems themselves should give us access to the origins and behavior of noise in quantum platforms on scales that haven’t been accessible before,” Kolkowitz says.
