The race to deliver quantum computing solutions that shield users from the underlying complexity of quantum computing is heating up quickly. One example is Multiverse Computing, a European company, which today launched the second financial services product in its Singularity product group. The new offering, Fair Price, “deliver[s] a higher accuracy in fair price calculations for financial institutions over current classical computing methods” according to the company.

This is a product, says the company, that’s now available and can be leased for 100,000 euros a year. Multiverse reports at least one bank is evaluating it for use in a ‘production’ environment.

It does seem as if the financial services sector will be an early adopter of quantum computing and so-called quantum-inspired classical computing solutions. Caveat: These are still very early days for all of quantum computing. Multiverse’s product announcement is occurring even as the fundamental technology ecosystem of quantum processors, quantum networking, quantum storage, firmware, etc., is still developing.

Multiverse’s broad idea is to deliver quantum computing solutions that match the problem tackled with the most appropriate underlying quantum technology. It’s a software company with extensive relationships within the quantum hardware community. The Fair Pricing product, for example, relies heavily on Monte Carlo simulation. For this application, Multiverse is leveraging IonQ’s trapped ion quantum computer. Portfolio optimization, the other product currently in the Singularity suite, has a different set of requirements. Multiverse uses D-Wave’s quantum annealing computer for this application.

Broadly, all of the quantum computing elements are hidden from the end user, explained Multiverse co-founder and CTO Sam Mugel during an interview with HPCwire. Half in jest, he said, “We’re quantum computing for the masses.” Indeed, the whole quantum computing community is working feverishly to develop tools to abstract away the need to know much about the underlying quantum technology.

All that’s needed to interface with Multiverse, says Mugel, is the ubiquitous Excel spreadsheet, a tool that is very familiar to quantitative financial analysts. Multiverse handles the under-the-hood quantum pieces. At least that’s the idea.

Formed in 2019, with a headcount of roughly 35 and growing, Multiverse is one of many start-ups seeking to build on quantum research that’s already done. They seek to act as a translating layer (for lack of a better term) to enable enterprises to use quantum computing. Like most quantum computing newcomers, Multiverse’s staff is dominated by Ph.D. mathematicians and quantum researchers. Mugel’s Ph.D., for example, was in cold atom quantum computing (link to founders).

Based in Spain, Multiverse is one of seven companies that recently formed a consortium (Amatech, BBVA, DAS Photonics, GMV, Multiverse computing, Qilimanjaro Quantum Tech y Repsol) and launched the CUCO Project to foster quantum computing research and development in Spain. The CUCO effort is another example of burgeoning state-funded efforts around the world; CUCO is aimed at implementing medium-term private-public quantum projects.

Mugel didn’t dig deeply into how the Fair Pricing product works. Other companies are leveraging quantum computing’s ability to quickly and effectively generate random numbers for use in cryptography and Monte Carlo simulation; presumably this plays a role in Multiverse’s approach. As described by Mugel, Multiverse has built a tool that takes input data (e.g. parameters) from the client via an Excel spreadsheet interface.

“The IonQ tool we built relies on quantum-accelerated Monte Carlo. This is an extremely hot topic at the moment and has stirred a lot of excitement in the finance community because Monte Carlo tools are so omnipresent. The motivation is that you have very well-known and well-defined bounds on how far you can push a Monte Carlo calculation. If you do so many samples, then you’re going to reach this accuracy. We know that statistically, and you won’t be able to exceed that classically, but you can go beyond that quantum mechanically,” said Mugel.

“We’ll run two calculations in parallel, or it’s an option to run two calculations in parallel. We’ve got default settings that will shoot [the job in the correct format] to IonQ. Then we’ll do a classical benchmarking in parallel and have a comparison of the two outputs essentially,” he said.

The company says using trapped ion quantum computers from IonQ with common PC-based software tools, the Singularity Fair Price solution can reduce error rates by 43 percent without increasing the number of runs or runtime. Below is a chart from a Multiverse paper, Quantum Portfolio Value Forecasting.

The portfolio optimization tool works differently.

“Monte Carlo can only be run on certain types of machines, which are universal quantum computers or gate model quantum computers. For optimization, we’re using analog quantum computing. Quantum annealing is a very specific type of architecture. What you’re really looking to do is solve an optimization problem and there’s are very efficient way of doing this quantum mechanically. We found there’s a very nice correspondence between the portfolio optimization tool, and the problem that’s natively solved by D-Wave. At the moment, we have a bank that has access to the code and is playing around with it and seeing if they want to put it into production,” said Mugel.

D-Wave, of course, is an early pioneer in analog quantum computing and a specialist in optimization problems. It’s also one of very few vendors who have actually ‘sold’ quantum systems as on-premise, stand-alone systems. Last fall, D-Wave expanded its efforts and started on a program to develop gate-based systems. The latter effort is in very early stages. (See HPCwire coverage, D-Wave Embraces Gate-Based Quantum Computing; Charts Path Forward)

Multiverse has published on its optimization work including this recent paper (Hybrid Quantum Investment Optimization with Minimal Holding Period) on arXiv on which Mugel was a lead author. Here’s the abstract:

“In this paper we propose a hybrid quantum-classical algorithm for dynamic portfolio optimization with minimal holding period. Our algorithm is based on sampling the near-optimal portfolios at each trading step using a quantum processor, and efficiently post-selecting to meet the minimal holding constraint. We found the optimal investment trajectory in a dataset of 50 assets spanning a one year trading period using the D-Wave 2000Q processor. Our method is remarkably efficient, and produces results much closer to the efficient frontier than typical portfolios. Moreover, we also show how our approach can easily produce trajectories adapted to different risk profiles, as typically offered in financial products. Our results are a clear example of how the combination of quantum and classical techniques can offer novel valuable tools to deal with real-life problems, beyond simple toy models, in current NISQ quantum processors.”

As you’d expect, Multiverse uses a blend of classical and quantum computing resources to obtain answers for collaborators and customers. Mugel declined to describe what specific types of classical system resources it uses, but did say much the work is done in the cloud. The figure below from the Multiverse website identifies its partners and they are certainly familiar names in the quantum community.

Core to Multiverse’s approach is matching problems with appropriate quantum technology; that’s an interesting challenge given steady advances being made among widely-varying qubit technologies. Like most quantum observers, Mugel expects there to be diversity of options.

“At least in the short- or medium-term, I believe that we’re going to continue seeing different platforms. For instance, take the Fair Pricing tool that we developed; we don’t actually need that many qubits for it. What we do need is very high fidelity and fully-connected architectures. It’s important that any qubit can talk to any other, and that we have many gates. We need very high [circuit] depths. Error rates really compound if you’re not careful. In this case, ion trap architectures were a godsend. But if we tried to apply ion trap architectures to portfolio optimization, they might not work as well. In portfolio optimization, we need lots of qubits, so that we can look at lots of assets. It doesn’t really matter if there’s lots of errors because even a suboptimal portfolio is actually already really quite interesting,” said Mugel.

The company is actively working with many qubit technologies.

“We’ve done lots of work with IBM’s architecture, also in the portfolio optimizations domain. We’re working a lot with ultra-cold atoms (companies). I’m really partial to this because this is what my PhD is in, and also it scales incredibly well. The problem is interesting: you have access to many more qubits, but the error rates are higher, but there’s also more connectivity. So potentially, these are more suited for certain types of problems like, for instance, hybrid calculations, like Variational calculations,” he said.

“We’ve got a very ambitious project ongoing with CUCO on how to solve quantum machine learning types of calculations on ultra-cold atoms. We’re also working with photonic architectures and are good friends and supporters of Xanadu. We’re looking at applying their architecture to energy markets. This is quite interesting. A client from completely outside of finance came to us and said, “Hey, we’ve got a problem that’s actually quite similar to portfolio optimization, we’re in a different field giving forecasts of supply and demands of energy. How do we optimize energy management?” And we said there might be a fit to solve this on like some injuries architecture was very well suited for complex network problems.”

With so much activity going on in the broad quantum computing ecosystem, it will be interesting to watch how companies like Multiverse fare in trying to deliver quantum computing’s advantage to users while hiding its complexity. While Multiverse’s current products serve FS niches, the company has ambitions to work with many industry segments.