Quantum Computing Use Cases for Financial Services

The large, fault-tolerant quantum computers of the future will have the potential to disrupt and revolutionize many aspects of the financial services industry. Some of the most commonly-cited quantum financial solutions include:

• Running Monte Carlo simulations with considerably more efficiency, thus enabling greater accuracy in financial forecasting
• Finding optimal mixes of various financial instruments, thus optimizing investment portfolios to minimize risks and maximize returns
• Analyzing significant volumes of financial data, thus allowing the identification of potential risks, as well as the discovery of suggestions as to how to mitigate and minimize these risks
• Calculating complex derivatives with speeds and accuracy well beyond the capabilities of classical computers
• Modeling data to find patterns, perform classifications, make predictions, and target customers in ways that are not otherwise possible with classical computing resources

Further commonly-cited use cases include:

• Assessing the credit worthiness of organizations and individuals, also known as credit scoring
• Determining the current fair market values of assets, also known as asset valuation
• Analyzing irregular behaviors, thus enabling fraud detection
• Aiding in the development of various trading strategies
• Predicting the likelihood and impact of negative events, also known as investment risk analysis

Again, the lists provided above are just the most common examples. Considerable investment and research into quantum computing for finance continues, and the discovery of other innovative uses can be expected.

Quantum Computing Demystified

Classical computers are built around transistors. Over the decades, these transistors have gotten smaller and smaller. They have gotten so small, in fact, that the physics we observe in the world around us begins to give way to the physics that governs the universe at the smallest of scales. This physics of the very small, known as quantum mechanics, interferes with the proper operation of these nanoscale transistors.

Rather than giving up due to this obstacle, two strategies have emerged. One, obviously, is to develop novel approaches to fabricating transistors. But the other is to try to actually harness quantum mechanics. And that’s where quantum computers come in.

Inside a quantum processor, there are no transistors. There may be some transistors nearby helping to control what’s going on, but the actual computation uses components much smaller than the smallest transistors. With one exception, the basic unit of quantum computation, called a qubit, is a particle. Common examples of such particles are atoms, electrons, and photons.

These particles can be so much more than the 0 or 1 that a transistor represents. In fact, the state of a single qubit can be represented by a complex number, which can encode quite a bit of information. But the real power of quantum computing comes from the interactions of qubits, which results in something called quantum entanglement. Whereas each additional transistor in a classical computer provides one more 0 or 1, each additional entangled qubit in a quantum computer doubles the amount of information that is encoded. Perhaps more incredibly, all of that quantum information can be processed simultaneously. Classical computers, in contract, process information sequentially.

While this may make it seem like quantum computers will solve all the world’s problems, that’s not actually the case. There are many tasks for which quantum computers underperform classical computers. However, there are problems that are extremely challenging, if not impossible, for classical computers. And it is with these challenging problems that quantum computers have the potential to excel.

A key takeaway here is that users need not know how quantum computers work at this level. As we note in our article titled “Exploring the Advantages of Cloud-Based Quantum Computing,” real quantum computers are available via the cloud. No knowledge of engineering, configuration, or maintenance is required to use them. This knowledge can be helpful, but it’s not required.

As we further note in our article titled “Quantum-as-a-Service: Definition, Advantages and Examples,” quantum applications are available via the cloud, as well. Applications of Quantum-as-a-Service include improved risk analysis, pricing models, and forecasting for the financial industry. Some of these services don’t even require any knowledge of computer programming.

Quantum Computing in Financial Services

Listing all of the various use cases in quantum computing in finance can seem repetitive, because many use cases are simply variations of the same approaches. The number of quantum algorithms and subroutines is still relatively small, but they can be applied in many different ways. We can abstract away much of the detail and provide a handful of high-level classifications:

• Risk analysis, which is the analysis of huge volumes of data with the goal being to identify potential risks so that it can then become possible to mitigate those risks
• Portfolio optimization, which involves finding the right mixes of financial instruments that allow the maximum possible returns on investments within acceptable tolerances for risk
• Fraud detection, which relies on the quick processing of vast amounts of data, thus making it possible to detect fraudulent activity in a timely manner
• Real-time trading, which again relies on the rapid processing of considerable volumes of information, and which enables decision making based on the latest market information
• Financial modeling, which involves the quick and efficient execution of complex calculations that are really slow, and potentially even impossible, with classical computation

This list of classifications will not grow as quickly as lists of use cases, but it is reasonable to assume that novel approaches to quantum computing in finance might require an expansion of this list from time to time.

Quantum Computing Use Cases in Finance

It’s one thing to merely list use cases and classifications. What’s more valuable is extracting analyses and insights from this information. A McKinsey & Company report titled “How quantum computing could change financial services” offers some such analyses:

• Although many finance problems are formulated as optimization problems, the earliest benefits of quantum computing might come simply from improving upon existing machine learning tasks
• The reversibility of quantum operations, regardless of the use case, may help financial companies satisfy specific legal requirements regarding the actions they take
• The scalability of the problem space, doubling with each additional entangled qubit, benefits all use cases, especially those that are resource-intensive
• Whereas classical solvers may have to restrict the dimensionality of data so as to not run out of memory, quantum solvers, in principle, have no such limitation
• Combinatorial problems are viewed as another early winner, especially with demonstrations of neutral atoms solving subsets of classically-intractable problems such as Max Independent Set
• Use cases, mentioned by name, include portfolio optimization, credit risk analysis, capital allocation, market risk analysis, smart routing, trade matching, private interbank trading, resource allocation, and tailored services

A Quantum Zeitgeist article titled “Quantum Computing in Finance: The Possible Use Cases” goes into greater detail into this first point. Examples of financial industry use cases are provided for regression, classification, clustering, generative modelling, feature extraction, reinforcement learning, and natural language processing machine learning tasks.

Finally, the second-to-last point needs to be made more often. Maximum Independent Set and Maximum Clique, for example, both have real-world use cases and cannot efficiently be solved with classical computers. However, their efficient solution has already been demonstrated on neutral atom quantum computers. All finance use cases that can be formulated as such problems, therefore, are poised to offer immediate competitive advantages as the number of atoms available scales upward.

‍