JPMorgan Chase has plenty of “quants” who hunt profits with computers. In 2018 the bank is adding employees you might call quantums. The computers they’ll use work on data using the intuition-defying processes of quantum mechanics.

America’s largest bank by assets is forming a small group of engineers and mathematicians to examine how quantum computers could help in areas such as trading or predicting financial risk. They’ll access supercooled quantum processors at IBM’s Yorktown Heights, New York, labs over the internet—a kind of quantum cloud. The bank is among a small group of companies tapping IBM’s prototypes, which if they can be scaled up should offer immense processing power.

Quantum processors built so far by IBM, Google, Intel, and some startups, are too small and unreliable to do useful work. But JPMorgan executives say they’re interested in part as a hedge against a future slowdown in improvements to conventional computers.

Chip experts say the phenomenon known as Moore’s Law that drove exponential gains in computing power for decades is now ending. Quantum computing could be a way to revive the rate of progress, at least in some areas. “If you can successfully apply it to problems it could give you an exponential increase in computing power that you can’t get” through traditional chip designs, says Bob Stolte, CTO for the equities division inside JPMorgan’s investment bank.

Microsoft CEO Satya Nadella cited the slowing of Moore’s Law as the motivation behind his own investment in quantum computing at the World Economic Forum in Davos last week. The company is working to build its first quantum computing hardware.

Microsoft grew to its current size by riding Moore’s Law, through close ties with the phenomenon’s primary driver, Intel. Now, Nadella said, times are not so good. “Moore’s Law is kinda running out of steam,” he told assembled financiers, government officials, and tycoons.

IBM Quantum Computing Scientists Hanhee Paik, left, and Sarah Sheldon examine the hardware inside an open dilution fridge at the IBM Q Lab.

Connie Zhou

The specter of a world in which silicon chips are not improving exponentially also helped convince Daimler CIO Jan Brecht to join IBM’s early-access program, dubbed the Q Network. “We don’t lack computing power today,” he says, “but you see Moore’s Law going into saturation.”

JPMorgan and Daimler’s quantum experiments may also have an element of keeping up with the Joneses. Other companies tapping IBM’s quantum processors include their competitors Honda and Barclays, as well as Samsung.

Another sign of growing interest: Startups working on designing chips, computers, or software for quantum computing received $248 million in funding last year, compared to just $43 million in 2016, according to CB Insights.

If and when they arrive, quantum computers won’t be good at everything. But physicists and computer scientists have proven, using theory, that even a relatively small quantum processor could do more than a phalanx of conventional supercomputers on some problems. Conventional computers work on data using bits that can be either 1 or 0. Quantum computers encode data into devices called qubits that can enter a “superposition” state in which they might be considered both 1 and 0 at the same time, allowing computational shortcuts.

As industrial interest in quantum computing has grown In recent years, researchers at Microsoft, IBM, and elsewhere have begun an intensive search for algorithms that could do useful work on even relatively simple quantum hardware.

Daimler’s Brecht says another reason he chose to work with IBM’s early prototypes is that quantum computers hold the prospect of being particularly useful at problems in which Daimler is interested, such as simulating chemical structures and reactions inside batteries. That could improve the range of eletric vehicles, or reduce their dependence on cobalt, a rare metal with a supply chain dogged by environmental and ethical problems.

Chemistry problems are probably the first that quantum computers will be able to help with, says Stephen Jordan, a quantum computing researcher who recently left the National Institute of Standards and Technology to join Microsoft. Simulating the quantum mechanics of molecules is a natural fit for qubits’ own quantum properties.

IBM’s quantum cloud currently consists of processors with 20 qubits, and the company has shown off larger versions with 50 qubits. Jordan says there’s “fairly good” evidence a 50-qubit processor, or not too much larger, could do useful work in chemistry if its qubits operate reliably enough. In September IBM showed it could simulate small molecules using just six qubits.

The path to tackling other problems on the wish lists of Daimler and JPMorgan is less clear. Brecht says the automaker also hopes quantum computers could optimize routes for delivery vehicles, or the movement of parts through factories. Some problems in finance, such as adjusting portfolio risk, can boil down to similar math.

So far, computer scientists haven’t proved what advantage quantum computers could offer on such problems, says Jordan. “We’re going to have to rely on experiments with future quantum hardware to figure out how advantageous quantum algorithms really are,” he says.

Dario Gil, the vice president leading IBM’s quantum project, says that’s exactly the kind of work his early-access program is designed to enable. “The mission is to discover the first quantum applications that have a commercial alignment,” he says. Gil and his partners in the auto and finance industries aren’t ready to say just when those money-making applications might properly arrive. “I definitely don’t have a timeline in mind, that’s not how we’re judging the success of this,” says JPMorgan’s Stolte. “This is very much exploratory research.”

Quantum quest



Source link

NO COMMENTS

LEAVE A REPLY