Google just hit a milestone in the development of quantum computers
The latest advancement, from researchers at Google and the University of California, Santa Barbara, is a way to make quantum computers more stable and better able to find and fix errors. The new research was published on March 4 in the journal Nature.
Some calculations take years for present-day computers to complete. A quantum computer could do that same calculation in a few seconds, and this breakthrough has some scientists saying that we're now half way to making quantum computers a reality.
What's different about a quantum computer
The key to quantum computers lies in qubits. The normal computers that we use every day use "bits" that store information as a 1 or a 0. A qubit, on the other hand, takes advantage of a really weird phenomenon in physics where tiny particles can exist in multiple places at once. That means qubits technically exist as both a 1 and a 0, so they can store exponentially more information than regular bits.
For example, a computer with two bits can encode information in only one of four possible combinations: 00, 01, 10, or 11. A qubit can hold all four of those combinations at once.
A quantum milestone
The trouble is that qubits are incredibly unstable and make a lot of errors. The information they hold only exists for a fraction of a second before their fragile state collapses. Even to exist for that fraction of a second, qubits require temperatures near absolute zero and an environment with no disturbances or noises.
And for quantum computers to ever work, scientists have to link many qubits together into arrays where the qubits are physically separate from each other, but all connected. So anything you do to one qubit instantly affects the others. That way the qubits can communicate with each other and quickly fix errors.
So the goal has become figuring out a way to stabilize qubits and to program them to find and fix any errors.
In a breakthrough, researchers from UCSB and Google have figured out how to stabilize an array of nine qubits.
While the array couldn't stop errors from happening in the first place, it could prevent the errors from messing up later calculations.
MIT physicist Scott Aaronsen pointed out in his blog that this experiment can be considered as completing 3.5 of the 7 steps needed to build a working quantum computer. In other words, we're half way there.
The experiment comes with an important caveat though: the errors that the nine-qubit array can protect against are only the kind of errors that show up in regular computers. Protecting against the errors that show up in quantum computers will require more complicated programming.
The researchers are optimistic though. Daniel Gottesman, a scientist who works with quantum error correction at the Perimeter Institute in Canada, told MIT Technology Review that he expects to see a demonstration of an array of qubits that can completely protect against quantum errors within the next few years.
Another important insight that this experiment revealed is that a group of qubits becomes more stable the more qubits you connect. That means that huge arrays made of thousands or even tens of thousands of qubits might actually yield a successful quantum computer some day.
John Martinis from UCSB, a key researcher in this experiment, has teamed up with Google to create the Quantum Artificial Intelligence Laboratory and they will continue experimenting. IBM has teamed up with Yale scientists and is also working on its own quantum computer.