Could College Park Be Ground Zero for a Quantum Computing Revolution?
To most of us, the key concepts of quantum computing might sound psychedelic. For instance, when atoms or other particles that store information realize (?!) they are being observed, they stop working and cause an error. But as long as no one peeks, the basic units of information in quantum computers can be in two states at once—both 0 and 1—a concept known as superposition. (In normal computers, binary bits of information must be one or the other.)
Then there’s this: Some respected physicists suspect quantum computers achieve uncanny power by tapping into computing resources that exist in … wait for it … parallel universes.
Here’s a secret: Scientists are still working to fully understand these concepts too. But that hasn’t stopped them from trying to lasso the weirdest parts of quantum mechanics—a branch of physics focused on nature at the atomic scale—to create technology with the potential to shake the world to its foundation.
A practical quantum computer, crunching some kinds of data exponentially faster than anything now available, could reinvent chemistry, allowing us to cure diseases with drugs now too complex to design. It might crack ironclad modern cryptography in hours, instead of the billions of years a standard computer would need to sort through every combination. It could begin to push artificial intelligence to disorienting heights.
Several global tech titans, various governments and a few smaller contenders have entered the race to build this superpowerful machine—and the University of Maryland has a hand in two of the most intriguing approaches, each with its own hurdles to overcome.
At one extreme, Microsoft is following a mindbendingly complicated, still-theoretical path known as “topological quantum computing,” with the help of Sankar Das Sarma, Richard E. Prange Chair in Physics. Then there’s the already operational “trapped ion” method driven primarily by Chris Monroe, Bice Zorn Professor of Physics, and relying heavily on exotic components like lasers and vacuum tubes.
“If there’s going to be a quantum computing revolution—and that is an ‘if,’” Monroe says, “it’s going to happen in College Park.”
Monroe (left) is entitled to make bold predictions. A pioneer in the field of quantum computing and a Distinguished University Professor, he’s the guy who made international headlines for “teleporting” quantum information from one atom to another via the strange but well-tested phenomenon of entanglement, spawning confusion (and perhaps travel plans) among those who took “Star Trek”-related terminology too literally.
Now he’s chief scientist and co-founder, along with Duke University engineering Professor Jungsang Kim, of IonQ, a startup just east of campus that has a few dozen employees, crucial technology developed by Monroe and licensed from UMD, and $20 million in venture capital to build a general purpose quantum computer.
It’s dwarfed by the big companies in the race, including IBM, Intel and Google, but the disparity isn’t quite as dramatic as it seems, thanks to IonQ’s connection to one of the premier quantum physics research enterprises.
More than a decade of partnership between UMD and the National Institute of Standards and Technology (NIST) has led to fundamental advances in quantum physics and information
science. Today, scores of physicists and a growing contingent of computer scientists and engineers work in a trio of research centers at UMD, run in collaboration with NIST: the Joint Quantum Institute, where Monroe is a fellow; the Joint Center for Quantum Information and Computer Science; and the recently announced Quantum Technology Center, where Monroe was recently named director.
Quantum computing is more than a business priority—it’s a national necessity, says Laurie Locascio, vice president for research at Maryland and a former NIST researcher who rose to the agency’s top levels.
“Both institutions have supported this effort because we saw it as a technological leading edge, and may be an example of a leapfrog technology for the United States that we can’t afford to be left behind on,” she says.
Private industry has always played a dominant role in computing, which is why UMD has been revisiting its policies on technology commercialization and licensing, as well as working to make Greater College Park an attractive setting for business, she says.
Monroe confirms that IonQ likely wouldn’t exist without UMD’s keen interest in fostering tech firms connected to the university.
“At Stanford or MIT, we’d probably still be waiting in line,” he says.
BUILT BY NATURE
Besides the NIST partnership, Monroe may have another advantage over the heavy hitters in the way the guts of his quantum computers work. They get their processing power from “trapped” atoms lined up by electrodes and suspended within a vacuum chamber. By zapping them individually with a laser, they can be manipulated to store information.
In a classical computer, the smallest units of information, or bits, are stored as 0 or 1. But thanks to superposition, quantum bits, or qubits, can be 0 and 1 simultaneously. With entanglement also in play, computer processing power increases exponentially.
Trapped ions—built by nature—are more stable than competitors’ manufactured qubits, and can avoid slumping into either a 0 or a 1, or “decohering,” for far longer. Delaying decoherence for seconds, as opposed to millionths of a second, keeps errors at bay.
“As memory elements, they’re perfect,” Monroe says of trapped ions. “Better than we really need.”
Last spring, Maryland researchers pitted his architecture against IBM’s, which relies on fabricated superconducting circuits to form qubits; each had five qubits. The IBM computer proved faster, but the Maryland entrant was more reliable and had more interconnected qubits. There was no clear-cut winner.
Since then, working in his labs in the Physical Sciences Complex basement, he and his research team have continued packing more ions into the vacuum tubes. They successfully used 53 ions—a record—in a quantum simulator (a specialty quantum computer designed for physics research) reported last fall in the journal Nature.
A new project—contained in black polycarbonate panels rather than spread out in a chaotic array of lasers and mirrors on a lab table—is the Monroe group’s first attempt at designing and building an integrated system. About a mile away at IonQ, a twin is being built as the company’s first prototype.
When fully operational, it will have 32 qubits in action. All would be under full control via laser, another new record.
How many qubits will a fully reliable quantum computer need—one with the power to accomplish things no classical computer can? Maybe a hundred thousand, Monroe says.
THE ERROR PROBLEM
Futuristic potential aside, compared to a $200 Costco special, today’s experimental quantum computers kind of stink, with comparatively astronomical rates of errors that result in lost information.
To get around this, scientists have developed systems for quantum error correction, in which secondary qubits back up and correct the main qubit when it goes awry. In classical computing, the same information can be stored in several copies, and if one bit goes bad, the majority with correct information will prevail. But because of those weird rules of physics, copying information on a quantum computer is the same as destroying it. Complicated workarounds are required.
The upshot: A single perfect “logical qubit” might require 1,000 imperfect qubits working together. A reliable, early-stage quantum computer able to outclass current machines might need 100,000 or more qubits, and even billions for ultra-powerful, world-altering computers.
But making error-free computing the priority this early in the game could be a mistake, Monroe says.
“Am I supposed to build a million qubits to get 10 perfect qubits? Anything I could do with 10 perfect qubits, I can already do on my laptop,” he says. “I’d rather have 100
imperfect qubits and look for something interesting I can do with them.
“It’s a subtle point, but if we don’t do that, we’re never going to get to a million qubits anyway.”
But what if it were possible to start with qubits impervious to errors?
That’s the idea theoretical physicist and Distinguished University Professor Sankar Das Sarma (left) and Microsoft are exploring through topological quantum computing, which he and two colleagues proposed in 2005. It’s named after the branch of math known as topology, which postulates that two objects as different as a doughnut and a coffee cup are essentially the same. Both have a single hole.
To simplify, if a qubit can be defined by its topology, or essential shape, minor errors shouldn’t hurt, Das Sarma says.
“Take a little bite out of a doughnut—that’s an error—it doesn’t change the doughnut topologically, and likewise, small errors don’t change the qubit fundamentally,” he says.
There’s much more to these “quantum doughnuts,” says Das Sarma: qubits composed of braided quasiparticles known as Majorana, recently discovered two-dimensional “emergent phenomena” that happen to be their own antiparticles, and … yeah, never mind. The point is, Majorana—whatever they are—might be able to form a completely stable qubit that doesn’t require a vast number of backups.
Whether the theory can become reality is an open question, though Das Sarma thinks it’s more likely to succeed than stashing millions of trapped ion qubits in vacuum chambers, or forcing perhaps an even-larger number of superconducting qubits to work together reliably.
“People who are big believers in the topological qubit believe that if a real quantum computer is built someday, it can only be a topological one,” Das Sarma says. “I probably belong to that category.”
There’s less circumspection from the company that has long helped bankroll Das Sarma’s theoretical journeys in the quantum world.
“We will have a commercially relevant computer—one that’s solving real problems—within five years,” Julie Love, Microsoft’s director of quantum computing business development, told the BBC earlier this year.
WORKING WITH IMPERFECTION
Engineering solutions to these issues, particularly quantum error correction, will take years of hard work, says Carl Williams, a physicist, quantum computing expert and acting director of NIST’s Physical Measurement Laboratory.
“To have a machine that’s big enough to do useful economic things—solving a problem in drug design or doing a major optimization of a key problem in aircraft routing—I really expect that is more than a decade off,” says Williams. “Some will say I’m too optimistic, some will say too pessimistic, but I believe we’re on a 10- to 20-year timeline.”
Until then, the world is going to have to live with quantum computers that may be the equivalent of the Wright brothers’ Flyer, if not Leonardo da Vinci’s fanciful flying machines.
John Preskill, a theoretical physicist at the California Institute of Technology (and a member of IonQ’s board of advisors), has coined an acronym to describe them: NISQs, or noisy intermediate-scale quantum computers. The devices just over the horizon won’t be perfect, and they’ll still leak quantum information like crazy, but they’re a necessary step.
“NISQ devices will be useful tools for exploring many-body quantum physics, and may have other useful applications, but the 100-bit quantum computer will not change the world right away. We should regard it as a significant step toward the more powerful quantum technologies of the future,” he writes.
IonQ aims to begin offering cloud quantum computing services soon, allowing users to log in and participate in what could be the dawn of a new technological age.
“We’re not the ones who are going to find out what this can do,” Monroe says. “We’re going to put this technology in the hands of people with problems they want to solve, and see what they can do with it. We’re going to learn from them.”
Philosophy Professor Has a Comic Take on Quantum Confusion
Distinguished University Professor Jeffrey Bub started out on course to write a book explaining quantum physics to non-physicists, but 2016’s “Bananaworld” only got halfway there before sinking beneath too many equations.
“I really wrote the book for my son, who at that point was a postdoc in physiology,” the physicist turned philosopher says. “But the math kind of got away from me and when I finished writing, it was pretty technical.”
It’s understandable, given the subject. Quantum theory, the foundation of many modern technologies, makes sense in mathematical terms. In everyday speech, it can sound deranged: effects without causes, or causes that follow effects; things that simultaneously exist in multiple states—one and two, red and blue, alive and dead—until you observe them and they settle into a single state. (But how does an atom know you looked at it?)
In a new book-length comic with a rigorous no-equations approach, “Totally Random: Why Nobody Understands Quantum Mechanics,” Bub takes another run at explaining quantum mysteries to his son—this time with the help of his daughter, Tanya Bub, his co-author and illustrator.
The Bubs don’t attempt to explain why the phenomena actually work (because no one knows), but do explore how they work in practice, using a simple metaphor of coin tosses and a liberal sprinkling of humor to examine the probabilistic nature of quantum reality, a concept that disturbed Einstein and prompted his famous declaration that “God does not play dice with the universe.”
Quantum physicists have mostly adhered to the maxim “shut up and calculate,” using quantum theory in their research but generally skirting the deep foundational questions that occupy Bub.
“Physicists often say, ‘You know, these are the kinds of questions you can discuss over a beer after you’ve finished your real work,’” he says.
So bring on the beer and hold the equations.
Imagining a Quantum Future
It might initially sound like computer science Professor Andrew Childs has one of the weirder jobs around: Essentially, he spends his days pondering, in complex mathematical terms, the amazing and important things you could do with a machine that doesn’t exist.
“I work mainly to identify problems that we could speed up solving by using quantum computers,” says Childs, an influential theorist in the field and co-director of the University of Maryland’s Joint Center for Quantum Information and Computer Science (QuICS). “I generally think about a setting in which we have a perfectly functioning quantum computer.”
To put it in aircraft terms, he’s thinking about the capabilities of a Boeing 787, when a wooden glider might be a better analogy to the current state of the art in quantum computers. To add a dash of wistfulness to the endeavor, the 40-year-old thinks such powerful quantum computers might still be in the works when he retires. (He’s more optimistic they’ll arrive within his lifetime.)
But nothing about his job is all that strange. The field of quantum computing took off in earnest after mathematician Peter Shor showed in 1994 that an entirely theoretical quantum computer far more advanced than anything that exists even today would solve prime factorization problems exponentially faster than a regular one.
It was the first demonstration that quantum computers could have real-world potential. (That factorization is the key to cracking modern cryptography didn’t hurt interest, either.)
It’s not that theorists like Shor or Childs were hunting for specific practical applications, such as codebreaking. They focus on fundamental quantum computing operations, and how to take advantage of strange aspects of physics like superposition and entanglement to increase processing speed and power—although it’s not a magic bullet for every type of problem, Childs warns.
But for appropriate ones, like simulating the immensely complex workings of a molecule for drug development, quantum computing could be transformative—just don’t hold your breath waiting.