As quantum computing technology continues to advance, financial institutions and researchers are investigating its potential applications in various facets of the financial industry, such as collateral optimization. However, the practical implementation of quantum computing solutions in collateral optimization will depend on its technological maturity and specific challenges and opportunities it presents to the financial industry.
HSBC, a banking and financial services organization, and Terra Quantum, a quantum technology company based in Switzerland, have announced a collaboration to explore hybrid quantum technology applications on optimization challenges for business impact.
In parallel, Terra Quantum has established a collaboration with Honda Research Institute Europe (HRI-EU) to create a solution for optimizing escape routes during disaster scenarios. In a world increasingly affected by climate change and extreme weather events, this collaboration could improve public safety.
Collateral optimization
The collaboration between HSBC and Terra Quantum aims to illustrate the possibility of a hybrid quantum solution for optimization challenges, which are a pervasive and very complex suite of problems in the financial services industry.
An example of this type of problem is known as collateral optimization—the process of effectively allocating and maintaining collateral assets to satisfy regulatory requirements and cut expenses as much as possible. To strike a healthy balance between risks, liquidity and profitability, it employs mathematical and computational procedures.
While existing methods for optimizing collateral generally rely on linear optimization solvers, which can occasionally find sub-optimal solutions when presented with higher levels of complexity, the quantum approach has the potential to outperform more conventional methodologies by successfully dealing with high-dimensional optimization issues and possibly having superior scalability.
“Our primary objective is to showcase the performance enhancement that is enabled through applying quantum algorithms to the collateral portfolio,” Terra Quantum founder and CEO Markus Pflitsch said in an interview with EE Times Europe. “Once we are successful, we would like to apply this approach at a greater scale, in the collateral portfolio but also to other optimization challenges.
“We can also apply our wider capabilities in quantum machine learning and simulation to other problems in financial services,” he added.
Terra Quantum’s initial approach entails reformulating the problem as a quadratic unconstrained integer optimization (QUIO) problem and applying its proprietary solver, TetraOpt, a search and optimization tool that can better handle higher dimensionality and non-linear constraints, thus optimizing and improving efficiency.
Hybrid quantum computing
Hybrid quantum technology refers to the integration of quantum components or systems with classical technologies to create more powerful and versatile computing, communication and sensing systems. This combination of classical and quantum technologies seeks to capitalize on the advantages of both while minimizing their respective weaknesses.
The main factors of hybrid quantum computing are the following:
Error correction: Due to factors like decoherence and noise, quantum computers (Figure 1) are susceptible to error. To improve the dependability and robustness of quantum computations, hybrid quantum systems frequently employ classical error-correction techniques, thus achieving low error rates.
Quantum algorithms: Hybrid quantum technology enables the development of algorithms that utilize quantum components for specific subroutines while relying on classical computing for the remainder of the computation, thereby accelerating the performance of certain tasks.
Simulation: Quantum simulation is possible with hybrid systems, in which a small quantum processor simulates the behavior of a much larger quantum system. The use of classical computers to control and read out the quantum processor enables the study of complex quantum phenomena.
Terra Quantum has a suite of algorithms in its portfolio that address a broad range of challenges in optimization, machine learning and simulation. These algorithms include:
Quantum encoding (QuEnc): A variational quantum optimization algorithm that uses exponentially fewer quantum hardware resources (qubits) to address challenges compared with common alternatives like QAOA
Hybrid quantum neural network (HQNN): Machine-learning algorithms that combine classical machine learning with quantum machine learning to enhance prediction accuracy and learning capacity
TetraPDE: Tensor-network–based simulations of complex systems, leading to significant speedup in solving partial differential equations
“These algorithms utilize or recreate quantum effects, such as superposition and entanglement, to address complex problems differently from classical algorithms,” Pflitsch said.
IBM Q Quantum Computer
As Terra Quantum explained, the complexity of collateral optimization is driven by the individual constraints that are imposed by each counterpart. For instance, each counterpart has specific requirements on asset eligibility, valuation haircuts, liquidity requirements, substitutions and so on. Additionally, taking into account factors like future derivative margining and collateral transformations makes this problem extremely challenging.
“Our approach with the TetraOpt quantum software tool is well-suited to dealing with problems with such complexity,” Pflitsch said.
Terra Quantum said its technology has demonstrated an advantage over traditional optimization methods in different use cases, including:
Enhanced option pricing with Cirdan Capital: Terra Quantum’s technology has increased the speed of exotic options pricing by 75%, reducing computing usage by 75%.
Satellite mission planning with Thales: Terra Quantum’s technology has improved solution optimality by 1.5%, leading to higher-quality image captures and enhanced revenues for the industry.
Biomass power plant optimization: By better predicting the steam mass flow, Terra Quantum’s technology has shown the potential to improve the operations of power plants, which can lead to significant reductions in excess CO2 and NO2.
Future perspective
While quantum technology is still in its early stages, Terra Quantum believes the industry is rapidly maturing. With the hybrid quantum approach, the company said it can unlock performance enhancement in certain problem types. This is typically done with quantum software, which can address complex problems executed on various classical computing infrastructures today.
“We are seeing rapid industry traction on the hybrid quantum computing approach today, and we are also seeing other companies in the quantum ecosystem following this path,” Pflitsch said. “As the quantum hardware matures over the next three to five years, we will see a significant enhancement in the performance of the applications developed and their impact on the financial industry and beyond.”
However, Pflitsch sees two major challenges that might hinder a wider adoption of this technology:
Quantum readiness: Organizations need to start getting quantum ready today, which involves upskilling the relevant stakeholders, identifying the right use cases and setting up the relevant infrastructure.
Active exploration: There will be no magic switch, and companies need to start exploring and developing applications with this technology today. If they don’t do that, then they will find themselves at a steep competitive disadvantage as the field matures. Many companies found this to be the case with the recent emergence of generative AI.
Terra Quantum is working with a wide range of players in financial services, from investment firms to global universal banks and hedge funds. Their hybrid quantum technology is also solving problems in industries like aerospace, logistics, automotive, chemicals and energy, the company said.
“We will continue our journey across algorithms, compute and security to deliver market-leading solutions that provide business-performance enhancement and security,” Pflitsch said.
Quantum technology meets real-world emergency scenarios
Disasters have become alarmingly frequent, with their occurrence increasing fivefold in the last five decades. Responding promptly to these crises is paramount, as it directly affects public safety. Yet the dynamic and unpredictable nature of such situations poses a formidable challenge for traditional technological solutions.
Terra Quantum and HRI-EU have explored the potential of hybrid quantum computing methods to optimize evacuation routes during emergencies. In a proof of concept, the teams have simulated an earthquake scenario on a realistic map of a small town. Partners said the solution has efficiently predicted dynamic escape routes for vehicles, significantly reducing evacuation times.
Terra Quantum and HRI-EU said the quantum computing solution can consider numerous real-time variables and has demonstrated competitive efficiency compared with traditional computing methods. It leverages quantum machine learning and requires access to only a fraction of the map’s information—less than 1%. This feature is crucial in the ever-evolving and uncertain context of emergency scenarios. The results are generated through quantum simulations performed on classical computing hardware, paving the way for potential future implementation on large-scale quantum computing hardware. For in-depth details on this collaborative work, refer to the recent preprint paper on arXiv. The future of disaster response could well be quantum computing.