What Quantum Computing Could Mean for Software Development

Last week, during Quantum.Tech Digital’s virtual Quantum Thursday sessions, a group of stakeholders from IBM, Harvard, Goldman Sachs, and the startup scene discussed scaling up and maturing quantum software development.

Talk of quantum computing -- the use of quantum mechanics to create compute power -- has been kicked around for some time. The hope is this will allow for exponentially faster and complex processing that outstrips current resources. The conversation last Thursday explored some aspirations tied to quantum software development, which would work with this caliber of computing, and what benefits industry might see from these efforts.

Quantum computing shares many characteristics with the emerging technology sector and is also an active field of fundamental research, said panel moderator, Sebastian Hassinger, quantum computing research and ecosystem partnerships global leader with IBM. “It’s a collaborative and iterative process,” he said. “There is a combination of researchers in industry and academia working together, often in very open science kinds of ways on the fundamental sides of quantum computing.” This is in addition to working with leaders in industry research who are looking at the potential applications of quantum technology, Hassinger said, particularly in areas where traditional computing might not have a path forward to meet certain critical challenges.

As speculative as quantum software development sounds, it is not an entirely alien concept. There is a broad class of quantum algorithms, said Yudong Cao, founder and CTO for startup Zapata Computing, that share similar features as machine learning models. “If you look at MLOps or AIOps, this is very much the sort of software engineering challenge [in quantum software] that people also face with AI.” He leads an effort at Zapata to provide software that might help industrial players explore possibilities of quantum computing.

Cao said when he started in the field, quantum computing was still largely an academic discipline with theoretical works that might predict what could be done with a quantum computer, as well as experimental works that demonstrate what could be done. “Today we’re seeing this gap become narrower and narrower,” he said. “On the theory side, we’re improving our algorithms to reduce the amount of resources. On the other side, new hardware is coming online.”

There is a frontier emerging for quantum computing thanks to software solutions and hardware maturing, but Cao said the confusing ecosystem needs to be sorted out. “What is needed is a set of tools that allow people to tap into this diverse landscape effectively,” he said. For instance, Cao said there can be software engineering issues such as framework compatibility, particularly when a developer wants to tap into tools from the open source world.

The scale of quantum computing could also introduce issues with data management, comparable to the intensive data seen with machine learning and AI, with the potential to scale up. “The dimensions of the data space can become very large,” Cao said.

Academic interests in quantum software development can be complementary in certain ways to industry’s need such computing resources, according to Prineha Narang, assistant professor at the John A. Paulson School of Engineering and Applied Sciences at Harvard University. She is also founder and CTO of startup Aliro Quantum.

Narang said in the academic arena, her research group thinks about quantum computing in terms of understanding correlations in quantum matter and how to predict them. “We want to be able to run really large calculations that tell us something that is new physics, new phenomena,” she said. “Something that is totally new and would not have been expected if you’d just done a classical or pen and paper calculation.”

On the business side, she said through her work at Aliro she sees industry focused on what quantum computing can do for them now. Industry is also concerned about what might happen with their work on programs if the quantum hardware changes or something more advanced comes along. “While these two sound totally different . . . what they are asking for is very similar,” Narang said. There is mutual demand in academia and industry for access to good quantum software, she said, that is reliable, and keeps up with the hardware. “They want to know they are getting the best performance,” Narang said. “They want abstraction but they don’t want so much abstraction that they lose performance.”

Overall, these are still early days for quantum software development, with new discoveries ahead, said Will Zeng, head of quantum research for Goldman Sachs. He said he has been assessing what affect quantum computing may have on core processes. “In the last few years, we’ve been able to show we can build some quantum computers as an industry,” Zeng said. “Not great quantum computers -- early, prototype quantum computers.”

Developing valuable applications with quantum is the big next hurdle, he said. His team works on resource estimation, which takes on problems that can be linked to a theoretical advantage for Goldman Sachs. A key task, Zeng said, is to think about prospects and significant advantages that might be realized if the right resources become available in this arena. “We would estimate how good of a quantum computer we would need in order to achieve that advantage,” he said.

Quantum software might eventually play a role in working more efficiently, Zeng said, but quantum computing still has a way to go before its potential is realized. “In the quantum space right now, we’re not ready to be planning production deployments yet,” he said. “We’re in the stage before that.”

For more content on quantum computing, follow up with these stories:

The Rise of TensorFlow

Is Quantum Computing Ready for Prime Time?

AI Explosion Brewing in the Cloud for 2020