A research team that includes two UO physicists have outlined new techniques
for controlling the building blocks of quantum computing, a potentially
significant step toward making such computers more accurate and useful.
Physicists David Allcock and David Wineland are founders of the new Oregon
Ions Laboratory, which was recently set up in the basement of Willamette
Hall. They are among 12 authors of a new paper, which is based on an
experiment at the National Institute for Standards and Technology in
Boulder, Colorado. Both scientists previously worked at the Colorado lab and
continued to collaborate on the project after coming to the UO in 2018.
The techniques, described in the journal Nature, involve the use of
trapped-ion quantum bits, or qubits, in quantum computing and simulations.
They could lead to improvements in the operation of quantum computers, which
still make too many computation errors to be effective tools, the physicists
said.
The problem with quantum computers is that their logic gates—the tools used
to perform basic logic functions in computing—"are really bad," Allcock
said.
"They fail about 1 percent of the time," he said. "You can do about 100
(operations), then you get garbage out."
Wineland added, "The whole field is in a stage now, because of errors that
exist, that we can't do lengthy calculations or simulations of practical
value on our machines."
The goal is to get to 10,000 operations without error and then add layers of
checks to fix the errors as they happen, he said.
"We want to get to that point," Allcock said. "Then you can use quantum
computers for something useful. Right now they're just toys."
Wineland said trapped ions are like a bowl of marbles that have certain
magnetic properties. Physicists can apply forces to the ions with different
methods, including lasers, Allcock said. But lasers are expensive and
complex machines, whereas making logic gates using magnetic forces is
cheaper and more practical because they can be generated directly with
integrated circuits, he said.
"What we did here is show these techniques work as well as anyone has done
logic gates before," he said.
Google and IBM are among the commercial enterprises that have armies of
engineers working on such problems, while academic physicists are trying to
show there are better, more basic techniques for solving them.
"We've shown you can do it in a technically simpler way," he said.
If physicists and engineers can make quantum computers reliable and able to
operate with large enough capacity, they could simulate other systems,
Wineland said. For example, a quantum computer could simulate the action of
a molecule used in drug therapy without having to synthesize it in a lab.
"There are some very practical, useful outcomes," Wineland said. "We're just
scratching the surface."
Reference:
Srinivas, R. et al. High-fidelity laser-free universal control of trapped
ion qubits. Nature (2021).
DOI: 10.1038/s41586-021-03809-4
Tags:
Physics