Scientists at IBM say they've developed a method to manage the unreliability inherent in quantum processors, possibly providing a long-awaited breakthrough toward making quantum computers as practical as conventional ones — or even moreso.

The advancement, detailed in a study published in the journal Nature, comes nearly four years after Google eagerly declared "quantum supremacy" when its scientists claimed they demonstrated that their quantum computer could outperform a classical one.

Though still a milestone, those claims of "quantum supremacy" didn't exactly pan out. Google's experiment was criticized as having no real world merit, and it wasn't long until other experiments demonstrated classical supercomputers could still outpace Google's.

IBM's researchers, though, sound confident that this time the gains are for real.

"We're entering this phase of quantum computing that I call utility," Jay Gambetta, an IBM Fellow and vice president of IBM Quantum Research, told The New York Times. "The era of utility."

At the risk of seriously dumbing down some marvelous, head-spinning science, here's a quick rundown on quantum computing.

Basically, it takes advantage of two principles of quantum mechanics. The first is superposition, the ability for a single particle, in this case quantum bits or qubits, to be in two separate states at the same time. Then there's entanglement, which enables two particles to share the same state simultaneously.

These spooky principles allow for a far smaller number of qubits to rival the processing power of regular bits, which can only be a binary one or zero. Sounds great, but at the quantum level, particles eerily exist at uncertain states, arising in a pesky randomness known as quantum noise.

Managing this noise is key to getting practical results from a quantum computer. A slight change in temperature, for example, could cause a qubit to change state or lose superposition.

This is where IBM's new work comes in. In the experiment, the company's researchers used a 127 qubit IBM Eagle processor to calculate what's known as an Ising model, simulating the behavior of 127 magnetic, quantum-sized particles in a magnetic field — a problem that has real-world value but, at that scale, is far too complicated for classical computers to solve.

To mitigate the quantum noise, the researchers, paradoxically, actually introduced more noise, and then precisely documented its effects on each part of the processor's circuit and the patterns that arose.

From there, the researchers could reliably extrapolate what the calculations would have looked like without noise at all. They call this process "error mitigation."

There's just one nagging problem. Since the calculations the IBM quantum processor performed were at such a complicated scale, a classical computer doing the same calculations would also run into uncertainties.

But because other experiments showed that their quantum processor produced more accurate results than a classical one when simulating a smaller, but still formidably complex Ising model, the researchers say there's a good chance their error-mitigated findings are correct.

"The level of agreement between the quantum and classical computations on such large problems was pretty surprising to me personally," co-author Andrew Eddins, a physicist at IBM Quantum, said in a lengthy company blog post. "Hopefully it’s impressive to everyone."

As promising as the findings are, it's "not obvious that they've achieved quantum supremacy here," co-author Michael Zaletel, a UC Berkley physicist, told the NYT.

Further experiments will need to corroborate that the IBM scientists' error mitigation techniques would not produce the same, or even better, results in a classic processor calculating the same problem.

In the meantime, the IBM scientists see their error mitigation as a stepping stone to an even more impressive process of error correction, which could be what finally ushers in an age of "quantum supremacy." We'll be watching.

More on quantum computing: Researchers Say They've Come Up With a Blueprint for Creating a Wormhole in a Lab

Share This Article