IonQ and Alice & Bob have separately announced breakthroughs that bring quantum computing closer to reality.
Alice & Bob announced that they improved error correction 160-fold by “squeezing” qubits together. And IonQ announced that they achieved an “orders of magnitude” increase in the speed of connections between different kinds of qubits.
Error correction is the holy grail for scaling quantum computers
Today, the problem with building usable quantum computers is that as you increase the number of qubits, the number of errors goes up even faster.
Quantum computing companies are pursuing multiple strategies for reducing errors in order to get them below the threshold level where they can hit, so to speak, escape velocity.
Alice & Bob’s latest research shows that the threshold for its kind of error correction will be easier to reach than previously thought, said Anil Murani, a senior researcher at the company
“And therefore, does bring us closer to escape velocity,” he added. “When this will happen, and we’ll be able to present an Alice & Bob’s logical qubit under threshold depends on a variety of factors, but we can openly say that this is the current big work happening in the lab.”
Alice & Bob uses cat qubits, which have built-in error correction. This is the sample approach that’s used by AWS’s recently announced Ocelot chip.
The new trick is “squeezing” the cat qubits, which reduces error rates and makes the qubits more stable, the company reported in a paper released earlier this month.
The new squeezed qubits aren’t yet available via the cloud for companies to test them out. However, the company’s existing cat qubits are available on Google Cloud via its Boson 4 chip.
There are three main paths quantum computing companies are following to reduce error rates low enough.
One is to use software to resolve errors. Another is to add redundancy — in effect, turning multiple physical qubits into a single logical one. The third approach, the one taken by Alice & Bob and several other companies, is to make the qubits themselves more stable and less error prone, said Heather West, research manager in the infrastructure systems, platforms, and technology group at IDG.
“The thought is that it’s going to be a combination of all three — and also maybe a different type of error correction that we don’t know yet,” she added.
The goal is to get the error rate down to the point that they no longer grow exponentially with every new qubit added.
“That’s the golden ticket to actually reaching a scalable quantum system,” West said.
And, so far, nobody’s hit that milestone, she says. Recent announcements are stepping stones, she said, but not yet the big breakthroughs we’re waiting for.
IonQ’s approach scales with specialized qubits
IonQ takes a different approach building a scalable quantum computer, using trapped ions — and claims that its new approach does, in fact, help it cross the critical milestone.
“We’re hitting escape velocity,” said Ricardo Viteri, staff physicist at IonQ. “This is a significant milestone.”
IonQ is now making gates faster while simultaneously making them more accurate — and doing so in a way that scales, he said. “This is more than a science experiment and more of a blueprint for the future of large-scale quantum computing.”
According to IonQ, by using the right kind of qubit at the right time, two-qubit operations can be performed at a rate of hundreds of nanoseconds — as much as 500 times faster than the current average.
The company tested combinations of different types of trapped ion cubits, including those using Barium, Calcium, Ytterbium, and Beryllium.
The trick, the company said in a technical blog post, was to use nanosecond-scale laser pulses.
The company is looking at how different types of qubits are better suited for different kinds of tasks.
“Leveraging mixed-species qubits enables specialized roles to be assigned to specific qubits, optimizing quantum systems,” said Michelle Joynson, an analyst at Juniper Research. The result is better computational efficiency, reduced errors from decoherence, and another step towards scalable quantum computers.
Another key aspect of IonQ’s announcement is the emphasis on communication between different qubits.
“The big application here is in implementation of quantum networking,” said Forrester Research analyst Brian Hopkins. “One of the big issues with scaling quantum computers is the fact that we can’t make different quantum chips work in parallel the way digital chips do.”
Instead, quantum states are translated into digital zeros and ones and sent over traditional networks, he said. “A better way to network quantum chips together will accelerate the race to quantum value.”
According to IonQ’s Viteri, the specific mixed-species qubits in this research are still theoretical, but the company’s existing Forte Enterprise system is already available on AWS, Azure, and Google Cloud.
Source:: Network World