When industry insiders talk about a future where quantum computers are capable of solving problems that classical, binary computers can’t, they’re referring to something called “quantum advantage.”

In order to achieve this advantage, quantum computers need to be stable enough to scale in size and capability. By-and-large, quantum computing experts believe the largest impediment to scalability in quantum computing systems is noise.

Related: Moody’s launches quantum-as-a-service platform for finance

The Harvard team’s research paper, titled “Logical quantum processor based on reconfigurable atom arrays,” describes a method by which quantum computing processes can be run with error-resistance and the ability to overcome noise.

Per the paper:

“These results herald the advent of early error-corrected quantum computation and chart a path toward large-scale logical processors.”

Noisy qubits

Insiders refer to the current state of quantum computing as the Noisy Intermediate-Scale Quantum (NISQ) era. This era is defined by quantum computers with less than 1,000 qubits (the quantum version of a computer bit) that are, by-and-large, “noisy.”

Noisy qubits are a problem because, in this case, it means they’re prone to faults and errors.

The Harvard team is claiming to have reached “early error-corrected quantum computations” that overcome noise at world-first scales. Judging by their paper, they haven’t reached full error-correction yet, however. At least not as most experts would likely view it.

Errors and measurements

Quantum computing is difficult because, unlike a classical computer bit, qubits basically lose their information when they’re measured. And the only way to know whether a given physical qubit has experienced an error in calculation is to measure it. Th

Full error-correction would entail the development of a quantum system capable of identifying and correcting errors as they pop up during the computational process. So far, these techniques have proven very hard to scale.

What the Harvard team’s processor does, rather than correct errors during calculations, is add a post-processing error-detection phase wherein erroneous results are identified and rejected.

This, according to the research, provides an entirely new and, perhaps, accelerated pathway for scaling quantum computers beyond the NISQ era and into the realm of quantum avantage.

While the work is promising, a DARPA press release indicated that at least an order of magnitude greater than the 48 logical qubits used in the team’s experiments will be needed to “solve any big problems envisioned for quantum computers.”

The researchers claim the techniques they’ve developed should be scalable to quantum systems with over 10,000 qubits.