Production Perig / Shutterstock

NSA, Army Seek Quantum Computers Less Prone to Error

Even ordinary computers flip a bit here and there, but their quantum cousins have a lot more ways to go wrong.

As the power and qubits in quantum computing systems increase, so does the need for cutting-edge capabilities to ascertain that they work. The Army Research Office and National Security Agency recently teamed up to solicit proposals for research that can help do exactly that.   

The entities launched a broad agency announcement this week to boost the development of innovative techniques and protocols that allow for Quantum Characterization, Verification, and Validation, or QCVV, of intermediate-scale quantum systems. QCVV is essentially the science of quantifying how well a quantum computer can run quantum algorithms—and experts agree that it’s a necessary step towards useful quantum computing.

“These new methods are sought as the next advances that will empower the quantum computing community to reliably interpret and evaluate emerging larger-scale quantum systems,” officials said in the solicitation. 

Traditional computers use inputs, instructions and outputs that are all sequences of 1s and 0s and individually called “bits,” but quantum computers use quantum bits, or “qubits.” The units of quantum information can take on other complex computing values outside of 0s and 1s, and offer an advanced set of calculation possibilities. 

Related: Don’t Rush Quantum-Proof Encryption, Warns NSA Research Director

Related: The Pentagon is Trying to Secure Its Networks Against Quantum Codebreakers

Related: The Pentagon Is Turning to Nature to Solve Its Most Complex Problems

Jacob Farinholt, a quantum scientist at Booz Allen Hamilton, told Nextgov this week that in classical computing systems, noise reveals itself as something “as simple as a bit flip, or maybe a lost bit here and there.” But one of the most powerful aspects of quantum computing is the fact that quantum resources grow exponentially with the number of qubits—so in quantum systems, there is also a continuum of possible errors that can occur. This means the slightest disturbance can propagate throughout the computation and rapidly destroy any information in the system. In order to engineer that noise out, experts first need to characterize it, hence the “C” in QCVV. Relatedly, verification and validation, or the “Vs,” refer to methods to assure that quantum computing systems were designed correctly and operate as expected. These challenges are equally as difficult, according to Farinholt. 

“Finding novel methods to characterize these noisy systems as they grow in size continues to be a significant challenge that will require some rather novel solutions,” he said. “Brute force evaluation of all possible computations with all possible input states is, of course, infeasible, so designing novel approaches to allow us to infer the information we need to perform verification and validation with finite resources is required.”

In the solicitation, ARO and NSA call for QCVV solutions for intermediate-scale “systems of size 10-20 qubits and larger systems greater than 20 qubits.” The insights could help pave the way for fault-tolerant quantum computing and systems that can better withstand negligible errors. The agencies plan to make awards that are subject to the availability of appropriations via procurement contracts, grants, cooperative agreements and other transaction for prototypes. Interested parties are invited to submit white papers on their potential proposals by Jan. 28, and formal proposals are due March 17. 

The agencies are accepting proposals across two categories in the space: integrated theoretical and experimental intermediate-scale QCVV, and novel theoretical approaches to intermediate-scale QCVV. While the first category engages research on a sort of intense collaboration between theory and experiment, the second hones in on the development of new scalable methods that will be implementable on much larger systems. “The two categories together indicate a clear and aggressive attempt to both conduct and rapidly transition this research into existing and future quantum computing technologies,” Farinholt noted. 

There’s no requirements for cost-sharing and ARO and NSA are seeking applicants from nonprofit organizations, small and large businesses, and foreign and domestic institutions of higher education. 

And though it’s early in the process, the community seems engaged—Nextgov spoke to professors at Canada’s University of Waterloo who are already preparing to ready-up their own submission. Quantum experts and UWaterloo professors Joseph Emerson and Joel Wallman explained that most QCVV research to date has primarily focused on characterizing one and two-qubit error mechanisms because of the sheer variety of error mechanisms in the systems and the fact that most experimental quantum systems currently in laboratories have one or two qubits. Wallman noted that quantum computers can only exhibit a significant computational advantage if there are more than 50 qubits, although the precise number is unknown. 

“The BAA is hoping to redirect the field to focus on larger systems where novel techniques are required to engineer and then characterize simpler noise mechanisms,” he said. “This focus on larger systems is crucial in the quest to build larger systems because there are many subtle effects that arise when controlling quantum systems—so we need good diagnostics to work out how to practically control such large scale quantum computers.”

In 2013, Wallman and Emerson responded to an earlier QCVV-related broad agency announcement from ARO—this one focused on “QCVV tools that will aid researchers as experiments begin to incorporate on the order of ten physical qubits.” They received funding and later co-founded a startup that sells software to implement quantum characterization methods, which they invented through their ARO-boosted research.

Emerson, who has over two decades of quantum-focused experience, noted that there are still some very significant deep-tech challenges “where we don’t know the path forward to get to truly disruptive quantum computing.” A great deal of quantum expertise is in academia and industry, so energizing efforts through solicitations like the new BAA could drive quantum advancements and stimulate economic impacts. 

“I think it’s great that we have interdisciplinary efforts through all these different layers of society,” Emerson said. 

Booz Allen’s Farinholt added that industry insiders are also excited to see the Army focusing on “such a critical area of quantum research.” He said federal support for quantum computing will prove to be critical for its ultimate implementation and applications to crucial, real-world problems. 

“Solicitations like this one that target a meaningful area in a highly technical way help galvanize scientists in the U.S. around key areas and help make us stronger together through focused effort,” Farinholt said. “The U.S. should continue to build on the quantum sciences momentum we have already accumulated—this will take place through continued federal support, strategic private investment, and academic study.”