A national intelligence R&D team wants to tackle problems too complex for classical computers, and thinks quantum technology could be the answer.
The Intelligence Advanced Research Projects Activity plans to award a multiyear grant to IBM to build out a small part of what could, in several years, make up a system capable of decoding encrypted information, or plotting the most efficient way to perform a series of tasks.
These tasks and others, such as breaking down the factors for very large numbers, prove difficult or impossible for computers relying on the traditional "bit" -- either ones or zeros -- while quantum computing relies on small units called “qubits” that could be in multiple states simultaneously.
David Moehring, program manager for this effort at IARPA, told Nextgov quantum computing might eventually be applied to challenges like the "Traveling Salesman Problem," which attempts to use algorithms find the most efficient way for the salesman to visit several disparate locations. It could also help analysts churn through large networks of databases, among other potential applications.
But it'll be several years before a true "quantum computer" is built. Right now, the agency is focusing on just one building block -- the "logical qubit" made of physical qubits -- that can remain unharmed against environmental influence and error, he said.
"The specific goals of this program are to get errors down to the 0.1 percent level," he added.
And work hasn't even begun, Moehring said. The research program, called "LogiQ," plans to hold an initial meeting early in the 2016 calendar year, and IARPA is still finalizing its contract with IBM. The agency plans to award multiple grants and contracts to other bidders in addition to IBM, though Moehring declined to share more details until negotiations end.
Moehring said the program is expected to last for about five years, but that IARPA's research programs are sometimes ended before completion depending on their progress.