Led by DSI members Henry Yuen and Sebastian Will, a new multi-university grant from the Air Force Office of Scientific Research will examine whether larger quantum operations could reduce errors and make future quantum computers more practical.
In theory, quantum computing has the potential to solve complex modeling problems that are currently beyond the reach of conventional machines, like simulating molecular behavior and designing new materials. Most existing quantum devices, however, implement calculations using the same sequential approach typical of traditional computing, conducting operations on one or two qubits at a time. Because of the uncertainty inherent in quantum states, these methods remain fragile, with errors often accumulating before calculations can fully execute. While many researchers are seeking to address these problems by improving hardware, error correction, or software designs, DSI member Henry Yuen is exploring ways to solve them by transforming the architecture of quantum computing itself.
As lead PI of a new, $7.5 million grant from the Air Force Office of Scientific Research, Yuen, the Srivani Family Associate Professor of Computer Science at Columbia Engineering, will lead a multidisciplinary team of researchers investigating whether one of quantum computing’s basic design conventions may be holding the field back. With administrative support from the Data Science Institute, the project will bring together experts from Columbia, Yale, and UC San Diego to comprehensively address the theoretical and practical parts of the problem.
Traditional computing carries out calculations using standard “bits”—digital switches that can either be on or off. While advances in the speed, power, and organization of traditional computers have made possible many impressive technologies, quantum computers promise a new frontier for both computing power and problem complexity. At the same time, today’s quantum computers carry out calculations using much the same approach as traditional computers, through long chains of small operations involving one or two qubits—the basic units of quantum information—at a time. In quantum computing, however, this sequential structure creates a compounding problem: every step introduces the possibility of error, and complex calculations may be overwhelmed by those errors before they are complete.
Yuen’s project represents a new approach: rather than relying on one- and two-qubit gate operations, his team will investigate whether using n-qubit gates–essentially, some larger number of simultaneous operations–could compress those long sequences of small steps into a single operation, thereby reducing the accumulated error. In one example from the proposal, a single 20-qubit fan-out gate could replace 690 sequential gates.
“You can think of computer code like a recipe,” says Yuen. “One philosophy is to break everything down into the smallest possible steps. But if there are instructions you use over and over again, it may be much more efficient for the computer to carry them out all in one go.”
While multi-qubit gates have been studied before, the approach Yuen’s team is taking is notable for testing the idea across the full quantum computing stack simultaneously: theory, compilation, error correction, and physical hardware. Running all four levels in parallel increases the likelihood of finding out early whether n-qubit gates are genuinely viable or just theoretically appealing.
To achieve this, Yuen and UC San Diego collaborator Daniel Grier will expand on the theoretical components of the approach, including the speedups these gates might enable, and the limits they may face. Yale’s Yongshan Ding and Shruti Puri will focus on compilation and error correction, translating abstract operations into machine-level instructions and managing the noise that disrupts real devices.
At Columbia, co-principal investigator Sebastian Will, also a member of DSI, will test whether these ideas can be implemented in neutral-atom systems, where lasers trap and manipulate individual atoms as qubits. By approaching these several facets of the problem simultaneously, the team can test and iterate across a range of possibilities efficiently—much like the n-qubit gates they hope to investigate.
Researchers hope quantum computers could eventually improve the simulation of chemical reactions, support materials discovery, and help answer scientific questions that remain beyond the reach of classical machines. But Yuen emphasizes that this research project is just one piece of a much larger challenge—one that depends on financial support for basic research.
“It’s incredibly important to have public and federal funding for this kind of work,” Yuen says. “It’ll take a while for the results to translate to what consumers see, but you definitely need it.”
Yuen credited Columbia’s Data Science Institute with helping make the award possible, citing DSI’s support throughout the proposal development process. The grant also reflects an important investment in Columbia’s wide-ranging work on quantum research.
For Yuen and his collaborators, this award also creates the space to pursue a much broader and potentially transformative question: not just how to build a quantum computer that works in theory, but one whose benefits can be realized in practice.