Ally Financial has partnered with quantum computing researchers to develop a new algorithm to enhance financial index tracking using quantum computing.
The digital financial services company worked with Multiverse Computing, a quantum computing solutions firm, and global consulting firm Protiviti to develop a method that optimizes investment portfolios automatically with returns that match traditional portfolios using significantly smaller sets of stocks.
Sathish Muthukrishnan, chief information, data and digital officer at Ally, said the Detroit-based bank, which has $186 billion in assets, has been experimenting with quantum computing use cases through its technology labs for more than a year.
“Quantum unlocks two things: the volume of data that I can use, and the concurrency of models that I can have,” Muthukrishnan said in an interview. “Executing different datasets together takes a lot of time using traditional models. Quantum brings the ability for me to do that at a much faster pace, giving me the opportunity to take scenario analysis to a completely different plane.”
Quantum computing uses qubits (quantum bits), which can exist in multidimensional states, to process data, while classical computers are restricted to storing information in a binary manner, either zeros or ones. Quantum computers can be used to sift through and analyze large amounts of numbers extremely quickly.
Ally has hired a few quantum physicists and partnered with institutions like Microsoft to research quantum computing on the tech giant’s Azure platform. For now, Muthukrishnan said the main goals are discovering use cases by partnering with other firms and companies. Ally doesn’t currently use quantum computing in practice, but the CIO said the company is in active conversations internally about how and when to implement the technology.
The recent research, conducted over a year, shows that the method, a hybrid classical-quantum approach, could build portfolios that outperformed the risk profile of a target index by up to two times, while using fewer stocks. The number of stocks in the researchers’ Nasdaq 100 fund was four times smaller than traditional portfolios, and ten times smaller in the S&P 500 fund.
In practice, financial institutions could use the algorithm to manage ETF funds with more speed and higher accuracy and returns, said Protiviti Director of Quantum Computing Services Konstantinos Karagiannis. Karagiannis said the notable aspect of this project, which was detailed in a recently released research paper, is the limited amount of stocks needed for the quantum computer to still outperform traditional portfolios. He added that while portfolio optimization can take up to 30 hours on some classical computers, the new algorithm runs almost instantaneously on a quantum computer.
“Right now we’re considering this customer advantage — this idea that, looking at what’s out there, we feel that this is faster, better, cheaper,” Karagiannis said. “For the first time we’re hoping to get with quantum what we’ve never had before…There’s all this trade off, we’re hoping quantum gives it all to you.”
There are headwinds to using quantum computing. Quantum computers aren’t commercially available, so companies use cloud-based quantum-computing-as-a-service offerings from companies like IBM, Microsoft and, in this case, D-Wave. Plus, only a handful of quantum computers on the cloud have the capabilities needed for major companies to run advanced programs, Karagiannis said, meaning gaining access can sometimes be a challenge. Developing algorithms that can run on quantum computers requires certain skills that not all quants and programmers have.
Still, Karagiannis said for applications that don’t need real-time solutions, there are worthwhile use cases for quantum. Muthukrishnan added that he wants Ally to lead the industry in quantum use when the technology becomes more mainstream.
“We’ll continue to experiment with quantum,” Muthukrishnan said. “We’ll continue to see how we can leverage our data using quantum computing, and see what value we can add for our business as well as our customers.”
Sam Palmer, head of financial engineering at Multiverse, co-authored the research paper with Karagiannis and Adam Florence, director of data science at Ally.