A scientific breakthrough lays the groundwork for powerful quantum computers that can tackle real-world problems in chemistry, medicine and materials science.
Using their Sycamore quantum processor, which is embedded in a super-chilled facility at Google’s facility in California to prevent interference from outside heat and light, the researchers completed a mathematical test in 200 seconds that they estimate the world’s fastest supercomputer would take 10,000 years to figure out.
The experiment used 53 “qubits”, the quantum equivalent of the digital ‘bit’ so integral to the computers we use every day. These qubits employ quantum principles to process ones and zeros much faster than conventional binary systems, because, in the quantum world, bits can also exist in states between one and zero.
Writing in the journal Nature this week, the researchers claim their experiment, which involved simultaneously producing one million samples of randomly generated numbers, achieved “quantum supremacy” – the point at which a quantum computer can do something that an ordinary or “classical” computer cannot.
It’s a claim that was immediately disputed by researchers at IBM, the other company throwing massive resources at fundamental research into quantum computing. I saw one of IBM’s quantum computers earlier in the year, complete with its steampunk tubes and cooling mechanisms designed to protect the delicate processors and quantum circuits from the noise of the outside environment.
IBM says that by employing vast amounts of memory and hard disk resources, its Summit supercomputer at Oak Ridge National Laboratory in Tennessee could have done the same job in 2.5 days.
“We urge the community to treat claims that, for the first time, a quantum computer did something that a classical computer cannot, with a large dose of scepticism due to the complicated nature of benchmarking an appropriate metric,” the researchers wrote.
Whatever the case, the experiment shows the immense processing potential that will be possible if scientists can maintain the low error rate the Google team achieved while adding more qubits to quantum computers.
“If Google, or someone else, upgraded from 53 to 55 qubits, that would apparently already be enough to exceed Summit’s 250-petabyte storage capacity,” wrote leading quantum computing expert Scott Aronson.
“At 60 qubits, you’d need 33 Summits. At 70 qubits, enough Summits to fill a city … you get the idea.”
There’s a reason why IBM, Google and others are so competitive in the quantum space. Not only does quantum supremacy represent a milestone up there with the invention of the transistor, which is in every digital gadget we use today, it also has huge commercial potential.
Take computational chemistry for instance. Scientists largely understood the chemical world by the 1920s, says Google’s Quantum Hardware Lead, Dr Anthony Megrant.
But coming up with new game-changing chemical compounds requires vastly complex computer modelling.
“The equations are much too complicated to actually solve for a lot of molecules and larger objects that are of interest to mankind,” he told a briefing of journalists yesterday.
“With quantum computers, we can use this vast resource of computation to simulate these large quantum systems. This can create better medicines, improve computational chemistry and lead to new areas of materials research.”
Megrant sees applying vast quantum computing power to produce cleaner energy as a key use that could benefit the world.
“We are facing growing energy consumption across the world. We can look at developing new catalysts for environmentally friendly plastics, or highly efficient solar cells for alternate sources of energy.”
A decade of work ahead
Scientists are already talking about developing a quantum computer with one million physical qubits undertaking such calculations.
But it is unclear how exactly how they will perform as they become more complex.
“The main challenges towards even more applications is we need to grow the number of qubits and reduce the errors associated with each qubit,” says Megrant.
“There's nothing fundamental stopping this scaling.”
The fragile nature of qubits means that they have to be carefully housed in facilities with dilution refrigerators at temperatures that get down to just above absolute zero. So they are unlikely to ever exist in the form of the desktop computers we currently use.
Instead, says Dr Andrew Dunsworth, who joined the Google AI Quantum Lab last year as a research scientist, we are likely to draw on quantum computers much as companies employ supercomputers to perform their calculations today – renting time on them, rather than owning them.
“You log in remotely, you put jobs on the supercomputer,” Dunsworth told NOTED.
Customers of a quantum computing service could hire quantum processing power to work on their most complex problems and supplement this power with classical computers to get the job done most cost-effectively, he says.
“It seems logical to think this is how general people will run quantum algorithms."
Getting to that point is at least a decade away, the Google scientists suggest. Huge technical barriers have to be overcome that will draw on fundamental research underway around the world, including here in New Zealand.
In Dunedin, the Dodds Wall Centre for Photonic and Quantum Technologies is looking at novel ways of creating qubits and developing solutions for quantum memory and debugging.
In the meantime, even that random number generation test used by Google to make its quantum supremacy claim holds promise. Random numbers are used for lotteries, in cryptography, security and to create computer simulations.
Creating true, certifiably random number samples could be incredibly useful for a range of applications.
But what about the worry that such quantum computers could undermine the immensely complex equations that underpin the encryption systems designed to keep our information and financial transactions safe?
“It is nowhere near capable of breaking cryptography that is in use today,” Dunsworth reassures us.
“It is way too small a quantum computer by many orders of magnitude to be able to do such a large computation.”
Still, Google and many others are working on post-quantum encryption in preparation for the day that quantum computers are powerful enough to decode the encryption systems we rely on in the digital world.
Ultimately, Google’s contested quantum supremacy is a symbolic milestone towards a future, still one to two decades away, where the exponentially greater computing power provided by the quirks of quantum mechanics open up all sorts of new potential applications – and a commercial battle for supremacy in the field.
“The analogue is the first flight of the Wright brothers back in 1903,” says Dr Yu Chen, another Quantum Hardware Lead at Google, who co-authored the Nature paper.
“Its flight was ten seconds. It was a big step for what it meant for the aviation industry. It was the beginning of that industry,” says Chen.
“If you look at what we have, quantum supremacy, we only performed this specific task, but for us it really proves, [the] quantum computing processor is working.”