Quantum computing, the use of the unique behaviours of subatomic particles as a basis for computer systems, has hogged the limelight in international debate about quantum technologies. From an investment and commercialisation standpoint, that’s unfortunate.
There has been a lot of noise about how entanglement (the inextricable link between pairs of particles, which Einstein referred to as “spooky action at a distance”) and superposition (how qubits can be at both 1 or 0 states simultaneously, unlike classical bits) may lead to a world of superfast, probabilistic computing. There’s a lot of noise within quantum computing systems, too, but that’s another matter.
A broad range of technologies
Yet quantum computing is far from the only star in the universe of quantum technologies, which has emerged from a physics that baffles and entrances even its own experts. Quantum technology embraces a host of different systems, each of which could form a fast-expanding sector of its own – if investors shift their focus away from computing. These include quantum timing, metrology, and navigation, such as the development of hyper-accurate, portable atomic clocks.
Quantum imaging is another, such as sensing how a single photon behaves in a complex environment, opening up a world in which cameras ‘see’ around corners. A related field is that of quantum or gravitational sensors. These can identify differences in gravity/mass below ground, and so are able to detect hidden substances or reveal voids and objects – a neat example of Einstein’s theories of the universe combining with those of quantum physicists.
Improved sensors, gyroscopes, and accelerometers could prove to be of benefit in applications such as autonomous vehicles of every kind, with GPS not functioning well in urban canyons – motorway cuttings or the streets of New York – or underwater, for example.
Quantum security and communications may offer even greater economic potential, in that they promise to be both fast and tamper proof – or at least, to reveal where and how any interference took place.
Where’s the tipping point?
But research has yet to reach a useful tipping point in these areas. Dr David Snelling, Fellow and Programme Director of Artificial Intelligence at Fujitsu, says:
Quantum communication, in some ways, isn't at a very technologically advanced state, but it works. I think what we need to look for in that space is the actual use case that we have urgently today that will drive business investment and business value from that technology.
If RSA was dead, where we would definitely have a business case today is quantum key distribution, but it's not there yet. So this is an area where we've got the technology in a pretty advanced state, but we don’t know what to do with it: the what for and the why.
Who holds the keys?
The challenge also remains that such systems still have to engage with the messy human world, plus the digital realm and the universe of big things. How to distribute quantum keys – especially at a distance? The answers to questions like that cannot be ‘spooky’; they will almost certainly demand the use of fallible classical systems.
For BT’s Professor Andrew Lord, Senior Manager of Optical Research, this is where his company comes in: as a provider of trusted nodes for nationwide quantum communications. Hopefully, the company will fare better in this space than it does in full-fibre broadband coverage – and that’s not to mention it regarding high-speed communications as a premium add-on service, rather than as a utility.
Nevertheless, all this diversity of opportunity is reflected in how one country, the UK, has responded by forming four academic research Hubs in a £120 million national programme. Those Hubs are: Sensors and Metrology; Quantum Enhanced Imaging (QuantIC); Networked Quantum Information Technologies (NQUIT); and Quantum Communications Technologies (aka the Quantum Communications Hub).
These are similar to the UK’s four national robotics Hubs, which share the aim of speeding ideas out of the research lab and into viable commercial products, though the robotics Hubs are organised more by application (nuclear decommissioning, space research, offshore engineering) than technology type.
As I explored in my companion report for diginomica/government, the UK’s strategic combination of quantum Hubs, central investment, venture capital, and the private sector is a model that other countries would like to emulate – a particular challenge across 50 US states, for example.
However, the sums of money being invested by the British government are not up to the challenges set out in its own Industrial Strategy. In short, the UK is pulling its well-targeted punches. Meanwhile, the fact that quantum is ‘hard tech’ – demanding patient capital and an initial focus on expensive hardware – is another obstacle to growing a quantum economy, for any country or organisation.
Solutions looking for problems
A third challenge – identified by Fujitsu’s Snelling – is that there are countless problems that industries such as healthcare, communications, manufacturing, transport, and logistics want to solve, but the tools don’t exist yet do that with quantum technology.
Ole Kock, Technical Authority in Quantum Sensors at industrial semiconductor specialist Teledyne e2v, says that his company is in listening mode to find out what customers actually want, suggesting that some aspects of quantum technology are indeed solutions looking for problems. The aims of business customers and quantum technologists are far from entangled, it seems.
At this point, it becomes apparent that quantum computing is not the UK’s particular focus: communications, security, and imaging are. Indeed, it was suggested by more than one speaker that quantum computing itself is likely to become a component of classical systems, either in simulations or simply within hardware accelerators.
Either way, quantum will not replace classical, said Dr Peter Waggett, Director of Emerging Technology at IBM Research UK. Instead, a wider ecosystem will emerge, with users’ technology choice depending entirely on application:
We need to move into spaces where we have different architectures for different jobs. It will be broken down into bits, which is standard high-performance computing, neurons, which is where we’re starting to develop architectures that apply to AI, and then quantum computing.
We know that the three of them need to be looked at together. And we need to make sure that we don't lose sight of the fact that there are some pretty spectacular things happening in standard computers.
However, he and other speakers accepted that further new challenges will emerge in this complex environment. One is that quantum security would make existing protocols obsolete, handing a significant advantage to the first country or organisation to deploy it at scale.
Another is that there is a development bottleneck between the classical and quantum computing worlds, according to Professor Elham Kashefi, Associate Director (Applications) at the UK’s Networked Quantum Information Technologies Hub (also Personal Chair in Quantum Computing at the University of Edinburgh).
Dr Claire Cramer, Embassy Science Fellow at the US Embassy in London, adds:
We don’t yet know what the truly transformative technologies will be. The likely best use cases will only be found through continued basic research. Time is also a barrier: a decade or more to implement transformative research in quantum. That’s a long time.
A question of transparency
Another barrier is more subtle: with AI black-boxes and neural networks already posing transparency, auditing, and explainability problems, the future emergence of quantum AI would make regulators’ and ethicists’ heads spin faster than an electron. The risk of technology reaching a level of extreme abstraction from the human world seems high.
Service giant Atos may, on the face of it, not be an obvious spokesperson for a technology that is speculative, amorphous, and (relatively) early stage, but Dr Crispin Keable, its Head of Big Data and High Performance Computing, explains:
We’re a services organisation, but we have a big R&D function and within that we do supercomputing. In supercomputing, the real challenge is that computers aren't getting faster fast enough – at least, not for the volume of data that users want to put through those machines, whether it's weather forecasting, or drug design.
So one of the motivations behind our quantum research programme is to try and change that, to develop quantum processors that will help in supercomputing. But if you think about what quantum computing is, compared to conventional computing, there are some very important differences: it's not deterministic, it's probabilistic.
The major differences between quantum and classical realms mean that algorithms and applications can’t simply be ported over from the digital world to run on quantum systems, which for programmers poses a problem. Keable says:
There has to be a way for people who are not specialists in quantum hardware to learn how to programme it. We’ve also heard a bit about noise [in quantum systems], and noise continues to be a problem. And so, because of that, the algorithm that you're developing has to work in a noisy environment. [...] Software development for quantum computing, to me, is the biggest issue.
One of the things that we expect is that most quantum computing systems will only be a part of a greater computing environment. [...] I think in the future is a situation where you can substitute the simulator for real hardware as it becomes available in a normal computer, and then you have a real speed-up for a particular application.
For Dr Carmen Palacios-Barraquero, young CEO of ambitious UK hardware startup Nu Quantum - more here - , the challenge there will the ratio between hardware and software startups, which is currently 10:90. She argues:
Hardware investors are really hard to find, because the hardware is expensive. So these investors need to be knowledgeable in tech, knowledgeable in hardware, and patient over long timescales.
This article and its companion over at diginomica/government expose the opportunities and good news about quantum technologies, but also the fault lines in the industry and the way innovation is nurtured and funded.
Most technologists and policymakers recognise that opportunities are there in terms of the big picture: a future that, if not quantum powered, will at least include quantum systems as accelerators, or as a means to see the unseen, or even to challenge the way that we perceive the world.
But ironically, it is less clear what those opportunities may be when it comes down to the details, the quantum level of investment. The closer we observe the opportunity, the more puzzling it gets. The challenge, then, is that most of us perceive the world via a mix of digital and analog systems, with AI, machine learning, and analytics exposing some of the underlying patterns. All of this complements our own human senses and ingenuity.
Put another way, in business and technology we see a world of big objects and quantifiable opportunities, and it is far from clear how the quantum realm relates to it – though it is clear that it does.
In short, investors, policymakers, and business leaders need something tangible and relatable before they reach for their credit cards. A use case, a killer application, a key to open the door. Until then, as I said in my previous report, the future will be both on and off at the same time.