Bits Blog: A Quantum Computing Problem Solved

University of Sydney Michael J. Biercuk, director of the Quantum Control Laboratory in the University of Sydney’s School of Physics.

Researchers at the University of Sydney and Dartmouth College said they have found a new way to design quantum memory, a key element in making quantum computing a reality. The method, they say, reduces the number of errors typically expected in quantum computing without sacrificing high-speed performance.

Quantum computing draws on certain well established, but difficult-to-control subatomic behaviors. In particular, it draws on the odd property of objects having both positive and negative charges at the same time, or for one subatomic particle affecting another without seemingly coming into physical contact with it.

Multiplied over many objects, this affords a great deal more computation than in the standard digital world of ones and zeros. Utilizing these properties could massively increase the power of certain types of computation, with implications for things like code breaking and security, materials science and facial recognition.

Compounding the difficulty of capturing and observing behavior in one of these machines, however, most quantum states exist for only the briefest fraction of a second. Keeping quantum information intact for long periods of time and keeping it relevant to the computation is one of the more daunting tasks in quantum computing. That is the problem the team says it may have solved.

The researchers used a technique called dynamical decoupling, or DD. DD has been shown to suppress errors in quantum systems by cancelling out fluctuations, something like the way one wave can smooth out a contending wave.

“The waves can overlap just the right way to build up a big amplitude or cancel out fluctuations,” Michael J. Biercuk, director of the Quantum Control Laboratory in the University of Sydney’s School of Physics, wrote in an e-mail. “In DD we need the interference to be just right such that the errors cancel.”

The team added to existing work in this field, he said, by figuring out how to break a sequence of behaviors into smaller segments that would preserve information without distorting the overall result.

“Amazingly, we were able to then show that even if we interrupted after some number of cycles, the error probability wouldn’t change much — meaning we could always ‘bound’ the error probability if we used repeated DD sequences” he said. “That is vital for a system designer who needs to know how his/her memory will perform.”

The results appear in the June 19 issue of the journal Nature Communications.

While the breakthrough is significant, Dr. Biercuk said, researchers must now show large-scale experimental demonstration of the process on a repeatable basis. The memory system must then be integrated with other error-correction algorithms to create more uniform results.

In addition to the computational uses, Dr. Biercuk said the results may be useful in the development of certain types of quantum-based communications technologies as well.