Standard Chartered keeps faith with quantum experimentation

Quantum computing technology is not yet a mainstay of the capital markets and it could be a decade or more before it is widely used. That said, some firms are trying to get ahead of the curve by discovering how they might integrate the technology into their existing toolbox.

Elena Strbac, global head of data science and innovation for corporate and investment banking at Standard Chartered, tells WatersTechnology that quantum computing is a complex area and building skills and knowledge takes time, which is spurring the bank to continue its experiments with the technology. 

“Our intention with all of our quantum initiatives is to future-proof ourselves and the service we provide to our clients by adopting new technology at an early stage,” says Strbac.

Standard Chartered is now focusing on building up its intellectual property and patent portfolio and sharing its learnings with key stakeholders, including clients, regulators, and governments it works with.

As a prerequisite, we are working closely with regulators to help inform what future quantum regulations might look like

Elena Strbac, Standard Chartered

One of its partners is the Universities Space Research Association (Usra), with which Standard Chartered has collaborated in quantum computing research since 2017. 

In 2020, the two signed a collaborative research agreement to partner on such research and develop quantum computing applications. The aim is to go beyond quantum annealing to include all unconventional computing systems that could provide an advantage to applications of interest, such as gate-model noisy-intermediate scale quantum (Nisq) processors and Coherent Ising machines.

A case for ESG

In 2022, the bank announced that it was looking at quantum-inspired machine learning for environmental, social, and governance (ESG) applications.

Strbac says the focus was on a specific ESG-related challenge: exploring whether quantum computing could improve the forecasting of natural disasters using satellite images. “Forecasting natural disasters is notoriously complex due to the chaotic nature of the systems involved,” she says. 

For both its quantum and classical approach, the team used a methodology called “reservoir computing”—a recurrent network with fixed random weights to extract the spatiotemporal information of the system—which showed promising results when trying to predict simple weather patterns, even with currently available quantum hardware. 

Its work uses a so-called hybrid quantum reservoir-computing framework, which replaces the reservoir with a quantum circuit. In its preprint article, the group proposed a HQRC algorithmic architecture that has a scalable modular structure and can be adjusted to accommodate hardware-efficient implementations. 

The paper describes standard reservoir computing algorithms and the hybrid quantum model—where the researchers explain the integration of quantum principles with classical reservoir computing methods to create quantum-enhanced learning models. 

It then presents a set of simulation results and a conclusion highlighting that the HQRC algorithm is an “interesting candidate” as a model for short-term forecasting of chaotic time series.  

The paper also notes the results are based on small system sizes, which can be easily simulated on a laptop. 

“Therefore, the proposed algorithm can also serve as a purely classical method. However, the simulation cost of full quantum circuits for the training and prediction phases is substantially more expensive than running state-of-art classical RC at this moment, so we do envision a hardware implementation,” it says. 

The collaboration prompted Standard Chartered to nominate Usra for The Earthshot Prize 2024, an award that recognizes the most innovative solutions to the greatest environmental challenges.

Future infrastructure

That project aside, Strbac says Standard Chartered has no immediate plans to put quantum computing initiatives into production. “As a prerequisite, we are working closely with regulators to help inform what future quantum regulations might look like, particularly regarding explainability; as well as working closely with our cybersecurity teams internally to understand how quantum computers can safely be part of our infrastructure in the future and how data would be transmitted securely,” she says.

The bank is now working with AWS, Imperial College London and Rigetti to explore how quantum computing can be used to improve classical machine learning techniques for analyzing complex data streams. This follows Rigetti securing an Innovate UK grant as part of the Feasibility Studies in Quantum Computing Applications competition. 

Combining quantum computing with classical machine learning could deliver more powerful tools for processing complex data. The consortium will use Rigetti’s quantum computer and software, Standard Chartered’s datasets and classical benchmarks, Imperial College London’s expertise in classical machine learning models for data streams, and AWS’s classical high-performance computing resources. 

Strbac says all datasets used in the project are public or open-sourced. The consortium will examine a number of predictive use cases and build classical and quantum machine learning solutions to solve the same problem in parallel. 

This is how it intends to create its “classical” benchmark. “We then run the equivalent quantum machine learning algorithm on a quantum simulator, and if the results are positive vs. the classical benchmark, we run it on the actual quantum hardware,” she says. 

The project will use Rigetti’s Ankaa 84-qubit quantum computer, which is hosted in the US, to do larger-scale experiments. 

Standard Chartered has been involved with Rigetti since 2020, when, together with Oxford Instruments NanoScience, the University of Edinburgh’s Quantum Software Lab and Phasecraft, it looked to develop the UK’s first quantum computer. 

Rigetti announced the project’s completion in April, resulting in a 32-qubit Aspen quantum computer. This is Rigetti’s first system to be deployed in the UK, and it was made available to the company’s UK partners through its Quantum Cloud Services platform.

Strbac explains that since partnering with Rigetti, Standard Chartered has identified certain use cases where quantum machine learning produces marginally more accurate predictions than its classical counterpart, and instances where quantum machine learning is able to solve the same problem with equivalent accuracy but with significantly less training data. 

“This was an interesting finding and a potential future differentiator for quantum,” she says. 

And then there were others

Not all banks are as experimental as Standard Chartered when it comes to quantum computing. Last year, for example, a former chief data officer at UBS said the bank had abandoned a multi-year effort to use the technology for trading after concluding it showed no “significant” advantage over existing tools. 

Lee Fulmer, who was chief data officer and head of the innovation lab at UBS until March 2023, said the bank ran Monte Carlo simulations against market data and volatility curves and was trying to use the technology to get an exponential “back out”—something that would give it a competitive advantage. 

Definitely, GenAI is one of those things that disrupted everything and took some of the attention from quantum, but there are still many opportunities

Julian van Velzen, Capgemini

The aim was to speed up the bank’s models, but Fulmer said UBS’s experiments failed to produce meaningful results. “In investment banking, you live and die by microseconds. The end result of all of that effort was that we found we weren’t getting a substantive uplift,” he said. 

When asked if he thought the technology had no real value in finance, Fulmer said, “Yes. If we look at how quantum computing started to evolve out of the 1990s, alongside blockchain, alongside virtual reality (VR), we have the same sorts of challenges with those. They are technical solutions for business problems that don’t exist.”

But others in the financial industry, such as Spanish banking group Caixa Bank, are still trialing the technology. In 2022, VidaCaixa, the group’s life insurance and pensions arm, developed two proofs of concept (PoC) using D-Wave’s Leap quantum cloud service and quantum hybrid solvers, which combine classical and quantum computing. 

CaixaBank’s two PoCs are for investment portfolio hedging and bond portfolio optimization. In the portfolio hedging use case, the PoC resulted in a 90% decrease in compute time compared with the traditional solution. For investment selection and allocation, the PoC delivered an application that optimizes the internal return rate by 10% in a chosen portfolio of bonds.

Julian van Velzen, chief technology and innovation officer and head of the Quantum Lab at technology services and consulting firm Capgemini, says quantum computing is maturing slowly, with attention shifting from the scale of qubits toward the quality of qubits. 

“A few years ago, there was a lot of focus on very wide circuits with low depths—so variational circuits like reservoir computing. Increasingly, we’re shifting away from this NISQ era to larger scale, fault-tolerance computers with more focus on the quality of the systems,” he says. 

This has resulted in more traction for companies prioritizing error correction and mitigation in quantum systems and the hardware providers building better systems. It’s a welcome trend, says van Velzen, as it will allow firms to run much deeper circuits and opens up many different use cases, such as derivatives pricing, risk analysis, and Monte Carlo-types of simulation for fault-tolerance regimes. 

“The issue, though, for these types of applications is that you need very deep circuits; you need billions of gates to run things. And with error correction, that makes it very slow. So, it will still take a very long time to solve it, and therefore it might not actually be providing any advantage. Additionally, as these machines are getting more mature, we are finding that even if there’s an advantage, in many cases there will be an economic disadvantage because it will be pretty expensive to run these things,” he says. 

This is somewhat in line with the conclusion of Standard Chartered’s pre-print article. Capgemini’s van Velzen adds that it helps explain why some banks have either stopped their quantum computing efforts or shifted their attention to other technologies. 

“Definitely, GenAI is one of those things that disrupted everything and took some of the attention from quantum, but there are still many opportunities,” he says. 

For example, van Velzen sees banks focusing on the field of post-quantum cryptography, also referred to as quantum-proof, quantum-safe, or quantum-resistant. 

He points to a letter issued by the Monetary Authority of Singapore urging financial institutions to keep updated with the latest developments in quantum computing and its associated cybersecurity risks. 

Though still relatively small, quantum-safe teams can now be found at most financial institutions. 

This is typically driven by CISOs, but increasingly, van Velzen says, CIOs are also getting involved. “So, it’s not just about the security, but also about the application infrastructure. And that’s interesting because it showcases the growing maturity and awareness that quantum-safe won’t be just a cryptography issue; it will be an application development issue, too,” he says.