Significance of Quantum Breakthrough Claim by Microsoft and Quantinuum

Yesterday (in early April 2024), Microsoft and Quantinuum announced a significant breakthrough in quantum error correction (QEC) for the trapped-ion quantum computer platform. The new QEC approach results in an improved reliability to generate a pair of qubits in Bell state (see my earlier post on how to create a Bell state), albeit with lower qubit resources used. Scott Aaronson also covered the significance of this announcement in his blog.

In short, there are actually two significant improvements achieved in this announcement.

Reduced Physical-to-Logical Qubit Ratio

According to their research paper, earlier works achieved using 10 noisy physical qubits to represent 1 reliable logical qubit. In this announcement, Microsoft applied its error-correction algorithm to Quantinuum's physical qubits, resulting in an improved ratio of using 31 physical qubits to represent up to 4 logical qubits. Microsoft claims this is the best ratio achieved, running over 14,000 experiments without error, significantly surpassing previous records.

Preselection instead of Postselection Bell Experiment

The second key improvement is the successful execution of a preselection Bell experiment instead of the insofar largely used postselection Bell experiment to demonstrate a successful quantum operation on a quantum computer. This means a higher assurance to execute a Bell circuit and consistently get expected Bell state outputs.

The key difference between preselection and postselection approaches to experiments lies in the timing of the selection process. Preselection involves choosing specific conditions or parameters before the experiment begins, guiding the setup to only produce outcomes that meet those criteria. In contrast, postselection occurs after the experiment has been conducted, where researchers select data or outcomes based on certain conditions or properties observed during the experiment. Preselection aims to control and manipulate variables beforehand, shaping the experiment's conditions from the start, while postselection allows researchers to filter and analyze data retrospectively, focusing on specific results of interest.

Conclusion

One thing to note is that the QEC approach used by Microsoft and Quantinuum is applicable specifically to the trapped-ion quantum computer architecture. It may not be feasible to other types of quantum computer architectures.

Nevertheless, the ability to deliver 4 logical qubits in a quantum computer may seem insignificant against the double-digit logical qubits supported in today's quantum simulators and quantum emulators. However, if the current pace of doubling reliable qubits in quantum computers continues, it is only a matter of time before quantum computers lead the logical qubit count race.