Microsoft and KPMG plan to test quantum computing with real-world problems

Alan Boyle

GeekWire


Krysta Svore at AAAS meeting

Krysta Svore, general manager of Microsoft Quantum, explains how quantum computing hardware works during a Seattle science conference in 2020. (GeekWire Photo / Alan Boyle)

Microsoft and KPMG are getting set to test Azure Quantum’s capabilities on the sorts of real-world problems that should give quantum computing an edge over traditional approaches.

Such problems have to do with optimizing systems and networks, such as where best to place cellular phone towers or how to allocate investments to match a client’s priorities relating to risks vs. rewards.

“Optimization problems are found in many industries and are often difficult to solve using traditional methods which can accelerate optimization,” Krysta Svore, general manager of Microsoft Quantum, explained today in a blog post. “Emulating these quantum effects on classical computers has led to the development of quantum-inspired optimization (QIO) algorithms that run on classical hardware.”

Such algorithms reflect the quantum perspective, in which information doesn’t necessarily take the form of rigid ones and zeroes but can instead reflect a range of values simultaneously during processing. The beauty of QIO algorithms is that they don’t need to run on honest-to-goodness quantum processors, which are still in their infancy.

The Microsoft-KPMG partnership gives both companies a chance to tweak the algorithms and how they’re used to maximize Azure Quantum’s QIO capabilities.

“The Azure Quantum platform allows us to explore numerous different solver approaches utilizing the same code, helping to minimize re-work and improve efficiency,” said Bent Dalager, global head of KPMG’s Quantum Hub. “The shared goal for these initial projects is to build solution blueprints for common industry optimization problems using Azure Quantum, which we can then provide to more clients at scale.”

This isn’t the first time for Microsoft’s quantum computing team has experimented with real-world optimization challenges: A couple of years ago, researchers at Microsoft and Ford used QIO algorithms to analyze strategies for smoothing out the Seattle area’s traffic snarls. Preliminary studies showed a decrease of more than 70% in congestion and an 8% reduction in average travel time.

Last year, Toyota Tsusho and a Japanese startup called Jij used Azure Quantum to optimize the timing of traffic signals. They found that QIO algorithms could reduce the waiting time for drivers stopped at red lights by about 20%, saving an average of about 5 seconds for each car. And California-based Trimble turned to Azure Quantum to identify the most efficient routes for fleets of vehicles, ensuring that fewer trucks run empty.

The Microsoft-KPMG project will start out focusing on benchmarking solutions for optimizing financial services portfolios and telecommunications operations.

Portfolio optimization has to do with balancing the mix of investments to minimize risk and maximize profit while staying within a given budget. As financial options get more complex, it becomes difficult to assess those options using a brute-force analytical approach — but QIO algorithms are well-suited to take on the challenge.

Quantum-inspired optimization could also increase the efficiency of voice-over-LTE telecom networks, leading to a better user experience for customers. Down the line, the project could look into optimization strategies for cell tower placement, mobile handover between cell towers, and staff scheduling for call centers.

“The teams plan to share results in the coming months,” Svore said.

Continue Reading

Loading data