Despite their enormous promise to solve new problems, today’s quantum computers are inherently error-prone. Scaling up invariably error-prone quantum processors is a formidable challenge. While quantum error correction ultimately promises fault-tolerant operation, the required qubit overhead and error thresholds are daunting.

Now scientists at the Pritzker School of Molecular Engineering (PME) at the University of Chicago have developed an entirely new technique for continuously monitoring the noise around a quantum system and adjusting the qubits in real time to reduce inaccuracy.

Their approach is based primarily on spectator qubits: a set of qubits embedded in the computer to measure external noise rather than store data. The noise in crucial data processing qubits can be eliminated using the information obtained from such qubits.

Assistant prof. Hannes Bernien, who led the research, said:With this approach we can very robustly improve the quality of the data qubits. I see this as very important in the context of quantum computing and quantum simulation.”

Theoretical physicists had previously proposed spectator qubits, which can be integrated into a quantum computer but do not store any necessary data. The spectator’s qubits would act like the microphone in noise-canceling headphones, tracking changes in the environment. Of course, a microphone can only detect sound waves, while the proposed spectator qubits would respond to any environmental changes that could affect qubits.

Bernien’s team wanted to show how this theoretical idea could be applied to reduce noise in a neutral atomic quantum array, which was their favorite quantum computer.

Atoms are held in place in a quantum processor with neutral atoms by laser beams called optical tweezers.

The first report on the creation of a hybrid atomic quantum processor with cesium and rubidium atoms was made in 2022 by Bernien and colleagues. They modified that processor so that the cesium atoms function as spectator qubits and the rubidium atoms as data qubits. The researchers created a system that continuously reads data from the rubidium atoms in real time and modifies the cesium atoms using microwave oscillations in response.

Bernie said, “The challenge was making sure the system was fast enough — any adjustments to the rubidium atoms had to happen almost instantaneously.”

“The exciting thing about this is that it not only minimizes the noise for the data qubits, but it’s also an example of interacting with a quantum system in real time.”

Bernien’s team subjected the quantum array to magnetic field noise to evaluate their error minimization method. They showed that the cesium atoms accurately detected this noise and that their system immediately wiped it out in the rubidium atoms.

The research team claims that the original prototype is just a starting point. They want to see if the method works by adjusting the types of disturbances and increasing the noise levels.

Bernie said, “We have exciting ideas on how to improve the sensitivity of this system by a large factor, but it will take more work to get it implemented. This was a great starting point.”

Ultimately, Bernien envisions a system of spectator qubits that could run constantly in the background of any neutral atom quantum computer as well as quantum computers of other architectures, minimizing error when the computer stores data and makes calculations.

Magazine reference:

  1. Singh et al. Mid-circuit correction of correlated phase errors using an array of spectator qubits. Science, 25 May 2023. DOI: 10.1126/science.ade5337