Google says error correction will lead to useful quantum computers

For the first time, Google has shown that it is possible to reduce the total number of errors generated by quantum computers. That means we should be able to build larger and more useful devices.

To Matthew Sparks

Google has demonstrated that its approach to quantum error correction, which is seen as a key part of developing useful quantum computers, is scalable, and its researchers believe that practical devices will be on the horizon for years to come. I’m sure you’ll be ready for it.

The building blocks of quantum computers are qubits, similar to transistors in traditional computer chips. But today’s qubits are susceptible to interference and errors, and these will need to be identified and fixed if we want to build a quantum computer large enough to really tackle real-world problems.

One common approach to this is called surface code correction, where many physical qubits act as one so-called logical qubit, essentially introducing redundancy. This is the degree of error correction capability in classical computers, but in quantum computers, each qubit exists in a mixed superposition of 0s and 1s, and any attempt to measure them directly destroys the data. so it gets even more complicated.

This means that adding physical qubits to logical qubits can actually be detrimental. “So far, the opposite has happened when engineers try to organize larger ensembles of physical qubits into logical qubits to achieve lower error rates,” he said of Google. Hartmut Neven says.

Google demonstrated this when it first announced a working error correction scheme in 2021. This resulted in a net increase in errors. Subsequent work at the Joint Quantum Laboratory in Maryland managed to reach a point where logical qubits did not exacerbate error rates, but at a technical level rather than a practical level.

Google now shows that the size of the logical qubit can be increased, and this scaling reduces the overall error rate. If this trend continues and quantum computers can scale, they will be able to do computations that even the most powerful classical computers cannot. Neven said there is “obvious confidence” within the team that Google will create a commercially useful quantum computer.

The team achieved the logical qubit milestone using Google’s third-generation Sycamore quantum processor with 53 qubits. A logical qubit in a surface code is typically a grid of qubits paired with another qubit of the same size, with a single qubit reserved for measuring other values. The company’s experiments saw a transition from a 3 x 3 grid containing 17 physical qubits to a 5 x 5 grid using 49 qubits. This meant that almost the entire processor was functioning as a single logical qubit. This increase reduced the error rate from 3.028% to 2.914%.

The Google team concedes the improvement is small, but says that in theory the scale-up process could continue indefinitely, paving the way for fault-tolerant quantum computers that can perform useful tasks reliably. I’m here. But moving to 6 x 6 logical qubits (including 71 physical qubits) is not possible on the company’s current generation of quantum processors and would require a major hardware step forward.

Fernando Gonzalez-Zalba of the University of Cambridge would have liked to have seen much better error rates, but says research is on the right track.

“As technology scales, improving logic error rates requires a little more improvement in the individual components within the processor,” he says. “[But] What you’ll see in the series of publications the team is producing is vast improvements from publication to publication. I don’t think it will be many years before we see scalable quantum error correction, but I think it’s pretty close. “