All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
Fintech is about more than offering customers funky colours for their debit cards. Several major banks – among them Barclays and JP Morgan Chase – are betting that the future of the industry is in quantum computing.
Both banks have begun to tinker with the technology by using IBM’s 20-qubit prototype quantum computer via the cloud, which is part of IBM's Q Network that gives companies and research labs access to several of the tech firm's quantum processors for research purposes.
While quantum computers are still in their infancy and have yet to outperform traditional computers, the two banks are already employing several physicists, mathematicians and computer programmers to prepare a breakthrough called "quantum advantage". As IBM defines it, that’s the moment when a quantum computer can run useful computing tasks better than a traditional computer.
On a standard computer, bits of data are in one of two states – on or off as represented by a 0 or 1. A quantum computer relies instead on qubits which can be one and zero simultaneously. It means that – in theory – a quantum computer can manipulate many combinations of data concurrently, leading to much more efficient data crunching. For that to work, these machines have to improve dramatically, because their error rates are still so high that they can’t outperform traditional computers at any task.
For financial firms, quantum computing is a hugely attractive concept. Back during the financial crisis of 2007-2008, poor risk assessment was a key troublemaker. Calculating risk – and understanding how different kinds of risk are related to each other – is an extremely complex job. But banks and hedge funds think that quantum computing might help them reduce the risk lurking in their investment portfolios.
Another big issue is the need to settle thousands of trades – matching up buyers and sellers with prices both sides agree on – at millisecond speed and high volume. The more efficient the settlement process, the better a bank can perform. “It's a very difficult problem classically,” says IBM mathematician Stefan Woerner.
To test whether quantum computing can help, a Barclays team of physicists and computer scientists under the helm of Lee Braine is busy writing quantum programs and running them on IBM’s machine. Ideally, says Braine, the transaction – the sale of a security like a share certificate and the cash payment – need to happen at exactly the same time. In quantum terms, this can be represented as a problem where “each transaction in a batch can be considered to be either settled or not settled – and the goal is to find the combination that results in the highest total number of transactions settled,” he says.
Woerner recently published a paper in the scientific journal Nature outlining how quantum algorithms could outperform traditional computers when it comes to analysing risk using what’s known as a Monte Carlo simulation. This simulation is traditionally used by financial institutions to determine the probability of an event while taking into account future risks.
During the last financial crisis, many banks and investors suddenly realised that their losses exceeded the money they had set aside for a rainy day. By running better Monte Carlo simulations on a quantum computer, they might be able to analyse many more financial scenarios and disasters in much more detail, allowing them to understand how likely it is that the maximum acceptable loss would be exceeded. Financial regulators could use such calculations to set the capital requirements for each bank with more confidence.
With traditional computers, running a comprehensive Monte Carlo simulation could take days to complete though, as the algorithms run through millions of scenarios. “A quantum computer can provide a quadratic speed-up – instead of many million scenarios, we only require a few thousands to achieve the same accuracy,” says Woerner – reducing the run time from overnight to nearly real-time. So instead of looking back and analysing the risk taken yesterday, a quantum computer would make it possible to react quickly to changing economic environments and make – or propose – decisions nearly instantly.
Right now, IBM’s quantum computers with their 20 to 50 qubits for computation are not powerful enough to do that, as to run a Monte Carlo simulation in real time would require thousands of qubits. “The banks need to be ready when the hardware is ready,” says Woerner. “So that when the hardware is available, they are there to realise this advantage if they want to be a first mover.”
Braine says that because of the limitations of current prototype quantum computers, Barclays has to “radically” simplify the business challenges when designing experiments for the quantum machine. But, he adds, he is looking forward to quantum processors having both more qubits and lower error rates.
This article was originally published by WIRED UK
