Quantum computing has spent years existing in the space between promise and practicality. The machines exist. Tech giants are building them. Researchers have shown that, for certain problems, quantum computers could one day dramatically outperform today’s fastest supercomputers. The obstacle isn’t that the computers don’t work. It’s that they’re extraordinarily fragile. Unlike traditional computers, which store information as ones and zeros, quantum computers perform calculations using quantum bits, or qubits. Those qubits are so sensitive that tiny interactions with their surroundings, such as imperfections in the hardware or subtle environmental disturbances, can introduce random errors into a calculation. Computer scientists call this quantum noise, and it remains one of the biggest barriers to making quantum computers practical. Baoyu Zhou’s work embraces the fact that quantum computers are imperfect. Instead of waiting for better hardware, he’s building mathematical tools that help researchers make better use of the hardware they already have. Zhou is an assistant professor of industrial engineering in the School of Computing and Augmented Intelligence, part of the Ira A. Fulton Schools of Engineering at Arizona State University. With a new three-year grant from the U.S. National Science Foundation, he’ll work with Xiu Yang, an associate professor at Lehigh University, on optimization algorithms designed specifically for today’s generation of quantum computers. The collaborative project will create scalable mathematical methods that continue performing reliably even when quantum hardware produces uncertain or noisy results. “The question is how we design more efficient, robust algorithms that can work despite the noise in today’s quantum hardware,” Zhou says. “I’m pretty optimistic that quantum computers will eventually become powerful enough for everyday use, and I want to help make that future possible.” Finding a signal in the noise Many of today’s most promising quantum computing techniques work through trial and error, repeatedly testing possible