Closer than you think – quantum computing in finance

May 2021  |  SPECIAL REPORT: BUSINESS STRATEGY & OPERATIONS

Financier Worldwide Magazine

May 2021 Issue


There are two competing narratives on quantum computing. The first is that the technology is overhyped, still in its infancy, beset by enormous technical challenges and that full-scale, reliable quantum computers are decades away. The second is that many companies, including some of the world’s most prominent financial institutions such as Barclays, BBVA, JPMorgan, Goldman Sachs and RBS are investing in, and experimenting with, quantum-computing applications and that business impact can be expected in the near term.

Which narrative is the correct one? Perhaps surprisingly, both are true. That creates an urgent need for finance leaders to understand the technology better in order to discern the short- versus medium-term implications for their companies, and to shape a viable quantum-computing strategy.

The information revolution of the last few decades was built on our ability to miniaturise transistors. In modern digital computers, now called classical computers to differentiate them from quantum computers, billions of little transistors switch on and off to execute binary logic. At the lowest level, a logical bit can either be a 1 (the transistor is on) or a 0 (the transistor is off).

Quantum bits, called qubits, are encoded and measured in two equivalent binary states called |0> or |1>, which are usually associated with the up or down spin of an electron. However, two strange properties of quantum mechanics, called superposition and entanglement, allow qubits to take on a continuum of non-binary values and interact with one another in that state to perform calculations. Any measurement collapses superposition, which means that the only two states a qubit can be measured in is either |0> or |1>.

While in superposition, qubits can be manipulated to process exponentially more information than classical bits. In fact, n qubits can process the same information as 2 to the power of n classical bits. When quantum computers eventually scale to hundreds of usable, error-corrected qubits, 2 to the power of n will become a truly astronomical number, allowing more calculations than classical computers can ever do. This explains the excitement about quantum computing technology.

The most popular qubit technologies are superconducting qubits used by companies such as IBM, Google, Alibaba, D-Wave and Rigetti, as well as trapped-ion technology used by IonQ and Honeywell. A third viable technology, photonic qubits, is being pursued by PsiQuantum and Xanadu. Other quantum technologies are nascent and largely unproven. Most quantum computing hardware need cryonic environments within a fraction of a degree of absolute zero (colder than outer space).

A basic quantum computer ‘stack’ comprises the quantum hardware at the bottom, with control systems above, topped by a software layer that compiles source code into executable programmes. A service level at the very top includes the operating system and software platforms that allow real-world problems to be translated into quantum-friendly algorithms.

So, where is the technology today and will it be ready for use in the short term? Being made from subatomic particles, qubits are extremely fragile and susceptible to mechanical, thermal or electromagnetic interference. When so affected, qubits lose their coherence, which is the largest source of errors in quantum computers. In order to properly compensate for decoherence, a substantial overhead of redundant qubits is required to perform error-correction. A fully error-corrected quantum computer will need to contain thousands of qubits to have enough qubits available for error correction. The technical challenges of handling and controlling that many qubits are at least a decade from being solved. The most advanced quantum computers currently have a few hundred qubits, insufficient for error correction. That means that we will have to live with high levels of errors for now. Today’s quantum computers are accordingly called noisy intermediate scale quantum (NISQ) devices. The NISQ era is expected to continue for the next five to 10 years at least.

Still the technology is advancing rapidly. In 2019, Google famously claimed ‘quantum supremacy’ after demonstrating that its quantum computer could perform a computation in minutes that would take the world’s most powerful classical supercomputer thousands of years. Since then, competitors and Google have made successive claims of having built the world’s most powerful quantum computer. New models are being released frequently by the various makers, each accompanied by fresh claims of superiority. While quantum supremacy means that a quantum computer can perform a task that a classical computer cannot, ‘quantum advantage’ is a more pragmatic term meaning that a quantum computer has an advantage over classical computers in performing a particular task, not that a classical computer could not do it at all.

Not many working quantum computers have been made yet, but the reach of the first quantum computers is vastly expanded by hardware makers allowing easy cloud access to the latest versions of their computers, sometimes even for free. Amazon, Alibaba and Microsoft are among the quantum cloud-service providers. A whole ecosystem of complementary platform and software solution providers has evolved, making it easy for anyone who wishes to experiment with quantum algorithms to run them over the cloud on their choice of hardware. No one needs to buy a quantum computer to use a quantum computer, which means that your competitors could be running proofs-of-concept for cutting-edge financial applications right now, without you even knowing about it.

Quantum computing promises to greatly speed up pharmaceutical and materials research by facilitating the simulation of thousands of molecules combined from atoms. The same ability to run an enormous number of combinations has application in the airline and logistics industries, where quantum computers can find optimal routes.

However, of all industries the finance industry has perhaps the most potential use cases for quantum computing. Several high-value problems in finance are in essence combinatorial optimisation problems, at which quantum computers excel. For example, pick a stock portfolio of 40 stocks from all listed stocks to maximise return and minimise volatility, or choose the optimal sequence in which to settle trades. Being able to test for an astronomical number of combinations also has applications in risk management, for example Monte-Carlo analyses could eventually be run in real time instead of in days. Combined with machine learning, quantum computing will accelerate the understanding of complex systems by machines, for instance to better find and target viable unbanked customers. In finance, unlike in most other industries, new algorithms can be deployed very quickly, which is why this industry is likely to see the earliest impact from quantum computing.

Researchers are learning how best to translate real-world finance problems such as asset portfolio optimisation into quantum computing algorithms. The current frontrunner technique for portfolio optimisation seems to be a class of algorithms known as quantum annealing. Annealing is a mathematical technique that finds the lowest energy state for a complex system. In the case of a portfolio optimisation problem, the problem can be constructed so that the lowest energy state corresponds to the highest return portfolio at lowest volatility. Quantum software start-ups Multiverse and Chicago Quantum, which both specialise in financial applications, have published encouraging portfolio optimisation results of such algorithms run on D-Wave’s quantum computer.

Such algorithms promise to solve what computer scientists call NP-hard problems in polynomial time rather than exponential time: in short, classical computers must sequentially compute billions of combinations for complex optimisation problems, which means the time needed grows exponentially with the number of potential variables until it takes even a supercomputer many years to do. However, quantum computers can simultaneously compute such combinations, meaning that the computation time is only a polynomial (such as a cube function) of the number of variables, which allows a really short runtime.

Currently-done brute-force calculations in asset management make ideal targets for quantum computing, with even small quantum computers already offering capabilities that classical computers cannot. Hybrid architectures, where classical computers shunt the data around and push only the part of the problem ripest for quantum processing to a quantum computer, enable annealing solutions to be deployed today.

The influence of quantum computing is not limited to software that runs on special quantum hardware. Quantum-inspired software uses quantum-like algorithms and mathematical models that can run on classical hardware with results that improve on traditional algorithms. For example, Fujitsu has built a quantum-inspired digital annealer using customised classical-computer hardware, while Toshiba offers quantum-inspired software that runs on standard classical computers.

There are now frequent media reports about financial services companies entering into partnerships with, and running experiments with, quantum computing providers. Quantum-computing providers and academics are publishing a growing body of papers that showcase their success at constructing optimal financial asset portfolios with historical stock-market data. It would therefore not be surprising if there were already asset managers exploiting insights from quantum algorithms to optimise their portfolios.

They have every incentive to keep quiet for as long as possible about the competitive advantage they gain from quantum computing. With cloud access, there is no technological barrier to entry in quantum computing. However, first movers are gaining an advantage in building expertise on how to model financial problems for quantum solutions, and by hiring already-scarce quantum talent. Even though full-scale quantum computers are still years away, it would be prudent for financial services companies to start gaining experience in the use of this nascent, but fast-moving, technology.

 

Peet van Biljon is the founder and chief executive of BMNP Strategies LLC. He can be contacted on +1 (202) 378 0294 or by email: ceo@bmnpstrategies.com.

© Financier Worldwide


©2001-2024 Financier Worldwide Ltd. All rights reserved. Any statements expressed on this website are understood to be general opinions and should not be relied upon as legal, financial or any other form of professional advice. Opinions expressed do not necessarily represent the views of the authors’ current or previous employers, or clients. The publisher, authors and authors' firms are not responsible for any loss third parties may suffer in connection with information or materials presented on this website, or use of any such information or materials by any third parties.