Making the Quantum Leap in Finance
Quantum computers used to be the stuff of sci-fi movies where futuristic robots and time machines transcended the realities of our imagination. Now the technology is no longer an illusion, and we are beginning to scratch beyond the surface of the possibilities of what it can bring to our lives — particularly towards how we approach finance.
So what does it mean to have a quantum advantage? We had a chat with Dr Mattia Fiorentini, Head of Artificial Intelligence of Cambridge Quantum Computing in the UK, about what this once-fictional technology is about and how it is serving the financial sector. Turn your chic geek mode on as we dive into the quantum realm with Dr Mattia.
Quantum computing seems like a complex subject. Can you help us make sense of it?
A quantum computer uses a fundamental set of principles of matter, which is quantum mechanics, to compute logical and mathematical operations. These deeper physical principles are not accessible to standard computers which makes quantum computers more powerful. The fundamental computational unit, the qubit, can be an atom, a particle of light, or a very cold, microscopical electrical circuit, as opposed to a silicon-based transistor (which is used in ordinary computers).
With this new technology, scientists now have the ability to easily solve problems that are, or have been, designed to be very hard to solve with an ordinary computer, such as designing new drugs, analysing large amounts of unstructured data, or even cracking passwords. We are now waiting for an engineer to build a quantum computer powerful enough to do all of this in practice, and more.
How does quantum computing help in innovating the way we currently conduct financial research and analysis?
Quantum computing can improve financial analysis with more articulated models and accelerate several scientific algorithms that are widely used in finance. As of today, it is possible to experiment with quantum algorithms for machine learning to increase the potential of data-driven modelling, optimisation to allocate funds in a portfolio, and Monte Carlo simulation for pricing financial instruments.
As an everyday consumer, what is the potential impact that quantum computing brings to banking?
Quantum computing will likely not be experienced directly by the end-user. Instead it could enhance some of the products that retail customers use every day. In the short term, the way credit card transactions and other kinds of highly sensitive data exchanges are made secure can be improved by quantum technology, also using qubits. As more powerful quantum computers become available, there is the chance that quantum machine learning models will improve all customer-related analytics to provide a more personalised experience.
How can innovators in fintech effectively use quantum computing to enhance their solutions?
This is will be highly dependent on the degree of automation and complexity of the technological solution that the innovators are adopting. The more “mathematical” or “scientific” the foundation of a process, the easier it is for quantum researchers to map that process to a quantum level and then elaborate a quantum-enhanced solution. Quantum computing services are currently delivered via cloud computing, so innovators should plan for this kind of IT infrastructure to host their services and product.
How will quantum computing impact AI and those who work on its development?
Quantum machine learning (QML) is a relatively small but vibrant area of study that can be implemented on current quantum hardware and hybrid quantum-classical cloud systems. Although it is premature to discuss if and how QML could help make progress towards artificial general intelligence, areas such as generative modelling can benefit from quantum models, such as the Quantum Circuit Born Machine, that could expand the capabilities of the field.
Interestingly, QML is also one of the areas of quantum computing that can take advantage of research ideas in deep learning. This means that ML practitioners can immediately contribute to developments in QML and, in particular, on topics such as novel optimisation method and quantum regularisation.
Can quantum computing be used to enhance cybersecurity? Conversely, will this pose a threat to our cybersecurity systems if it falls into the wrong hands?
Quantum systems, and qubits in particular, can generate streams of random bits that are certified to be random. This can be readily used in current encryption protocols to generate more resistant cryptographic keys.
On the other hand, we have all heard of the Shor algorithm that promises to factor integer number in a very short time, rendering many widespread commercial-grade cryptographic protocols obsolete. Fortunately, cracking cryptography with quantum computers is a task that will be out of reach of available quantum hardware for at least a decade or more. Also, quantum-resistant cryptographic schemes are in the process of being standardised for adoption by the wider public.
How do you see quantum computing developing in the future?
Quantum computing will become a very valuable technology for researchers in many fields such as physics, chemistry, computer science and the related applied domains. As our strategies for manufacturing Quantum Processing Units develop, quantum computing will not replace the current High-Performance Computing paradigm but complement it, and increasingly attract funding from the IT sector.
Scientifically, we will strive to find other experiments to demonstrate quantum advantage. In the years to come, specific technological improvements will be desirable, such as the integration of Quantum Processing Units with Quantum Random Access Memory: if built, QRAM could render quantum computing a game-changing technology in the big data and machine learning field. As much as these technologies still seem a distance away, experts believe that the current state of the art should be a source of optimism, and the technological breakthroughs to make this happen could be closer than expected.
Dr Mattia Fiorentini is Head of Machine Learning and Quantum Algorithms at Cambridge Quantum Computing (CQC) and is a Doctor of Philosophy in Physics from King’s College London. Dr Mattia leads CQC’s research team in high-impact applications of quantum algorithms in the fields of machine learning and optimisation whose work includes the development of deep learning to financial time-series modelling and decision making.
