Quantum computing has a number of potential applications in finance and it is drawing more and more attention, both in the financial sector and among researchers.
Fifty years ago, few people could have imagined that we would have microcomputers in our pockets with the same processing power as the computer that put a man on the moon. However, today, smartphones are ubiquitous, and as time has passed technology has assumed an essential role in all of life’s spheres, from telecommunication to science. But despite these developments, there are still problems that a regular computer cannot solve.
There are many challenges that we would like to solve, but the lack of processing power makes it difficult. During the financial crisis that shocked the world in 2008, poor risk management was a major problem. A quantum computer would allow the financial sector to conduct discoveries in a much shorter time span.
Maybe you have heard of the binary system that computers use. A classic bit can either be a 1 or a 0. On or off. But a quantum bit can have both values at the same time, and quantum bits can work together.
As we build more powerful quantum computers, you will be able to make more risk calculations in less time than you can today
Jan B. Lillelund, CTO, Executive Architect at IBM
“As we build more powerful quantum computers, you will be able to make more risk calculations in less time than you can today. Within a period of five to ten years we will be able to make greater risk assessments. In ten years you will then be able to make more precise and comprehensive analyses that give a better picture of how the financial systems work together,” says Jan B. Lillelund, CTO, Executive Architect at IBM.
Quantum computers scale in a different way to ordinary computers. Each time you add a qubit you double its computing power, whereas in a traditional computer you have to double the number of transistors to double the computing power.
Quantum algorithms outperform traditional computers
The Monte Carlo simulation is the standard method for dealing with systems with intrinsic randomness. Financial institutions use the method to determine the probability of an event and account for future risks. Because of the ‘unknowns’ in the system, it is impossible to make a prediction of how the system or process will evolve, so what is done instead is to statically sample realisations of the system.
“The precision of the simulated result can be increased by increasing the number of samples used, and if the distribution from which the samples are drawn is wide (large uncertainties in the system) it becomes computationally hard. However, quantum algorithms exist, offering a quadratic speed-up in the calculation of expectation values by Monte Carlo sampling. Thus, a future quantum computer with sufficiently numbered and sufficiently high-quality qubits would be able to outperform present supercomputers in risk analysis based on Monte Carlo simulation,” says Ulrich Busk Hoff, postdoc at DTU Physics.
Foreseen applications for quantum computing in the financial sector are portfolio optimisation, finding arbitrage opportunities, performing credit scoring, and risk analysis. The common denominator for those applications is that they deal with uncertainties – uncertainties due to incomplete knowledge about the market. Because of this, the realm of financial prediction has an intrinsic randomness to it that can only be dealt with by statistical analysis. The race is then about making these analyses a bit faster and a bit more detailed than the competitor’s. Ultimately, this is a race on computing power.
“Generally, the problems in finance are optimisation and belong to the class of computational problems that cannot be handled efficiently on a traditional computer. That means that the computational time required scales exponentially with the size of the input. In some cases, the problems can in principle be solved on a traditional computer, but the time would take unfeasibly long – even for the most powerful super computers,” says Hoff.
The advantage of quantum computing is that allows us to transform some of those hard problems into tractable problems. One of the first, and most well-known, such quantum algorithms is “Shor’s algorithm”, which turns prime factorization into a tractable problem. The difficulty of prime factorisation is the underlying principle for RSA encryption, which is widely used for secure online communication.
Quantum computers are not expected to replace classic ones. They will almost certainly require classic computers for interfacing and programming. It is unlikely that quantum computers are going to be in ordinary homes in the future. Rather, they will be located at nodes of a quantum network where they can be accessed remotely – much like the Quantum Experience already made available by IBM.