Europe-based quantum computing pioneers Multiverse Computing and Pasqal, and global bank Crédit Agricole CIB today announced successful conclusion of a 1.5-year POC study “to evaluate the contribution of an algorithmic approach inspired by quantum computing, and the potential of quantum computers, in two areas: the valuation of financial products, and the assessment of credit risks.”

The collaborators are positioning their work as a significant step forward for the use of quantum computing in financial services. Pasqal is developing quantum computers based on neutral atom qubit technology. Multiverse specializes in quantum software and algorithms. Crédit Agricole is one of France’s largest banks.

“These two Proofs of Concept demonstrated the potential and reality of quantum computing for finance, despite these technologies still being in their infancy. We took advantage of this initiative to start developing the internal skills to prepare for a technological breakthrough which, if it happens, will have a direct and decisive impact on competitiveness in our sector,” said Ali El Hamidi of Crédit Agricole in the announcement.

Georges-Olivier Reymond, Pasqal president, is quoted, “This is the most instructive experiment carried out in the industry so far, offering concrete comparisons for the first time, launching a new era for quantum computing. One of the results is that the tipping point is not that far away, probably less than two years, and that it is therefore urgent for users to quickly adopt these new methods, as Crédit Agricole CIB has done.”

Financial services has been among the more aggressive sectors seeking to understand and deploy quantum computing for competitive advantage. (See HPCwire article, JPMorgan Chase Bets Big on Quantum Computing). The joint study tackled three questions:

• Derivatives calculation. “To assess the performance gain offered by quantum computing in the valuation of derivatives. Recent research has shown the benefit of neural networks for this type of calculation. Yet, in several cases, these neural networks are difficult to use because they are too resource intensive in terms of memory and suffer from lengthy processing times. However, algorithmic techniques inspired by quantum computing can be used to optimize the speed and memory required for this training phase, leading to faster valuations and more accurate risk assessments.”
• Counterparties downgrades anticipation. “The goal of this experiment was twofold: first, to measure a quantum computer’s ability to solve a concrete problem, given the current state of technology. Second, to assess the change in performance depending on the number of qubits used. The bank chose a production use case, providing a real point of comparison: the anticipation of a counterparty credit rating downgrade over a 6 to 15-month period. Through conventional computer technology and heuristics, good results can be achieved. However, these methods do not work for all problems, and there is no guarantee that the results obtained will be close to the ideal solution. Using quantum parallelism, in theory, makes it possible to find optimum solutions more efficiently.”

The collaborators say both experiments were successful: “A marked improvement in computing time requiring a smaller memory footprint was measured using quantum computing techniques, paving the way for their use in real-world applications in the valuation of derivatives. For the quantum computer, the chosen problem was tackled under real-world conditions. With a quantum processor of only 50 qubits, the results obtained are as accurate as the results in production. Our projections indicate that this performance could be bettered at 300 qubits, a power that should be available industrially in 2024.”

The companies cite papers supporting the work, which provide more granular detail. The most recent was posted on ArXiv in January (Quantum-Inspired Tensor Neural Networks for Option Pricing). Here’s the abstract:

“Recent advances in deep learning have enabled us to address the curse of dimensionality (COD) by solving problems in higher dimensions. A subset of such approaches of addressing the COD has led us to solving high-dimensional PDEs. This has resulted in opening doors to solving a variety of real-world problems ranging from mathematical finance to stochastic control for industrial applications. Although feasible, these deep learning methods are still constrained by training time and memory.

“Tackling these shortcomings, Tensor Neural Networks (TNN) demonstrate that they can provide significant parameter savings while attaining the same accuracy as compared to the classical Dense Neural Network (DNN). In addition, we also show how TNN can be trained faster than DNN for the same accuracy. Besides TNN, we also introduce Tensor Network Initializer (TNN Init), a weight initialization scheme that leads to faster convergence with smaller variance for an equivalent parameter count as compared to a DNN. We benchmark TNN and TNN Init by applying them to solve the parabolic PDE associated with the Heston model, which is widely used in financial pricing theory.”

A second paper (Financial Risk Management on a Neutral Atom Quantum Processor), posted in December, showcased the ability of Pasqal’s neutral atom-based quantum computer. Here’s its abstract:

“Machine Learning models capable of handling the large datasets collected in the financial world can often become black boxes expensive to run. The quantum computing paradigm suggests new optimization techniques, that combined with classical algorithms, may deliver competitive, faster and more interpretable models. “In this work we propose a quantum-enhanced machine learning solution for the prediction of credit rating downgrades, also known as fallen-angels forecasting in the financial risk management field. We implement this solution on a neutral atom Quantum Processing Unit with up to 60 qubits on a real-life dataset. We report competitive performances against the state-of-the-art Random Forest benchmark whilst our model achieves better interpretability and comparable training times. We examine how to improve performance in the near-term validating our ideas with Tensor Networks-based numerical simulations.”

The authors say their work represents the first quantum-enhanced machine learning solution for the prediction of credit rating downgrades, also known as fallen-angels forecasting.

“Our algorithm comprises a hybrid classical-quantum classification model based on QBoost, tested on a neutral atom quantum platform and benchmarked against Random Forest, one of the state-of-the-art classical machine learning techniques used in the Finance industry. We report that the proposed classifier trained on QPU achieved competitive performance with 27.9% precision against the benchmarked 28% precision for the same recall of approximately 83%,” say the study authors in the paper’s conclusion.

“However, the proposed approach outperformed its classical counterpart with respect to interpretability with only 50 learners employed versus 1200 for the Random Forest and comparable runtimes. These results were obtained leveraging the hardware-tailored Random Graph Sampling method to optimize QUBOs up to size 60. The RGS method showed similar performances with simulated annealing approach and was able to provide solutions to QUBO within acceptable repetitions budget.”

Many observers expect 2023 to be a year in which many similar POC efforts and several early production use cases reach fruition.

Link to third paper cited in the work by Pascal, Multiverse, and Crédit Agricole (Quantum-Inspired Tensor Neural Networks for Partial Differential Equations), posted last August, https://arxiv.org/pdf/2208.02235.pdf.