Career and personal growth at SAP Labs Waterloo
The David Johnston Research + Technology Park is a community where challenges are tackled, advancements are made and new ways…
A new tech hub is emerging that promises to revolutionize the lives of future generations in the way Silicon Valley has transformed ours. Where is this place? Closer than you think.
Raymond Laflamme almost dropped out of graduate school back in 1983. He was a talented young physicist, but the program at the University of Cambridge was unforgiving. Hunched over his books day and night, he remembers being “invisible” to even his closest friends. But the intensive studying paid off and once his grades were posted — a very public affair at Cambridge — a classmate approached him with a message: “Professor Stephen Hawking would like to see you.”
Laflamme would spend his first year with Hawking engrossed in a particularly daunting problem. What would happen to the flow of time if the universe stopped expanding and began to contract? Hawking expected time itself to run in reverse, “but I kept getting something different,” recalls Laflamme. Each time they met, Hawking would suggest a slightly different approach — but no matter what Laflamme did with the equations, he kept getting the same answer: time kept marching forward. In the end, Hawking admitted he had made a mistake (a confession immortalized on page 150 of A Brief History of Time, for those who made it that far).
After years spent pondering the universe, Laflamme turned his attention to the very tiniest scales: the quantum realm. What is the nature of the quantum world? Can we learn how to manipulate information using the rules of quantum mechanics? The work that had occupied Laflamme at Cambridge was both ambitious and profound — space, time, the universe and all that — but it won’t get you a better smartphone. Quantum computers, on the other hand, could be a game changer. “We’ve found a fundamentally new type of machine,” he says. “It can do things that we couldn’t imagine were possible before.”
Born and raised in Quebec City, Laflamme, now 54, lives in Waterloo, Ont., where he serves as executive director of the Institute for Quantum Computing (IQC), headquartered on the campus of the University of Waterloo. The institute was launched a dozen years ago, the brainchild of Mike Lazaridis, billionaire founder of BlackBerry (formerly Research in Motion). Recognizing the potential of quantum information science, Lazaridis teamed up with David Johnston, at that time president of the University of Waterloo, to establish an institute dedicated to probing the basic science behind this new field. (A few years earlier, Lazaridis had founded the Perimeter Institute for Theoretical Physics, also located in Waterloo.) With funding from the Canadian and Ontario governments, the University of Waterloo, an array of corporate partners and a good chunk of Lazaridis’ personal fortune, IQC was soon up and running. Laflamme, who had been working at Los Alamos National Laboratory in New Mexico, was recruited as director. In 2012, the facility expanded into a new, 285,000-sq.-ft., $160-million building known as the Mike & Ophelia Lazaridis Quantum-Nano Centre.
Researchers at IQC are leading Canada’s foray into the quantum world, but they’re not the only ones peering eagerly into the subatomic realm. Some of the world’s largest corporations, from Google to Lockheed Martin, are getting in on the action. So is the US government (according to information leaked by Edward Snowden last year, the US National Security Agency has put US$80 million into a quantum computation project called Penetrating Hard Targets), as well as the UK, the European Union, Switzerland, Australia and Japan.
To see why so many are banking so much on the idea of a quantum computer, we need to embark on a brief tour through quantumland. In the everyday world of trees and cars and baseballs and houses — what physicists call the “classical” world — you can usually tell where something is and where it’s going. But zoom in by a factor of a few billion and the picture changes. At the quantum scale, a particle can be in two places at once — or, more generally, it can be in two “states” at once — a phenomenon known as “superposition.” And that goes for the fundamental “bits” used in computers, too. In classical computing, a bit can be either a zero or a one. But a quantum bit, known as a qubit, can be both a zero and a one at the same time.
Thanks to superposition, qubits can be used to perform huge numbers of calculations simultaneously. Moreover, because the power of a quantum computer scales exponentially in proportion to the number of qubits in its memory (two qubits can perform four calculations at once, three can do eight, four can do 16, and so on), it offers an exponential increase in computing power over today’s machines. “If we had a small quantum computer, with 40 to 50 qubits, we could do calculations that, if we had all the classical computers on earth, we would never be able to do,” says Laflamme. “It’s a mind-bogglingly powerful parallelism.”
One of the most-often touted applications is in cryptography. Much of the data that whizzes through the Internet today — everything from credit card purchases to medical, business and government records — is secured using a method called RSA coding. At the heart of the system is a simple truth about multiplication and division. Suppose you take two prime numbers that are hundreds of digits each and multiply them together. That’s easy; any computer can work out the product. But given only the product, working backward to deduce the two prime factors is a nightmare; no ordinary computer is up to the task. A quantum computer, on the other hand, could handle it — effectively rendering today’s best codes obsolete. At the same time, quantum theory has some researchers thinking about new kinds of codes — perhaps unbreakable ones. Other potential applications range from drug development and medical imaging to database management and artificial intelligence.
The path from blackboard to laboratory, however, is proving to be a long one. For starters, it’s not immediately clear what a qubit should consist of. In classical computing, the usual choice is electrons — or rather, pulses of many thousands of electrons; each electron carries an electrical charge and manipulating bundles of electrons is fairly straightforward. Engineers learned how to design “logic gates” — tiny electrical switches — for manipulating those bundles of electrons, allowing for the construction of circuits that can implement algorithms. The first logic gates used vacuum tubes; later these gave way to transistors. But qubits need to be in a state of superposition, which makes everything a lot harder. Photons (particles of light) are one possibility; ions (atoms that carry an electrical charge) are another; superconductors (supercooled materials, usually metals that carry electrical currents without any loss) are yet another. “It may take us a while to figure out which is the best architecture,” says Aephraim Steinberg, a physicist at the University of Toronto.
Another problem is that quantum states are inherently fragile; if you poke at them even slightly they “decohere” — that is, they stop displaying quantum properties and start acting like ordinary classical matter. But you need to poke at them to input and output information — in other words, for computation. Yet another problem is error correction. Each time a bit is processed there’s a risk of making a mistake. Classical computers get around the problem with redundancy; instead of doing something once they do it multiple times to make sure they keep getting the same result. Quantum computers can get by with fewer bits — because each quantum bit is so powerful — but at the same time, each qubit is far more prone to error. But progress is being made, says Laflamme. “We’ve gone from a few qubits with about 10% error 10 years ago to 0.01% error and a dozen qubits today,” he says. “It’s progress we can quantify.”
While IQC’s current focus is on pure science — discovery for its own sake — Laflamme insists there will be a payoff. “What we want to do in the next 10 years is to take advantage of this science, to turn it into technology, to commercialize it,” he says. Quantum computation could revolutionize any field where huge amounts of data need to be processed. Lockheed Martin, for example, is using the technology to test the software used to run its jet aircraft. Transportation experts say it could vastly improve analyses of traffic patterns in the air and on the ground. And Google is said to be using quantum computers to help its self-driving cars cope with the massive quantity of real-time data that must be processed each second in order to navigate safely.
Other potential breakthroughs, still under investigation, tug at the imagination. Take quantum teleportation, for example. Although the science is light-years away from the transporters that made Star Trek so compelling (“Beam me up, Scotty”), last year scientists in the Netherlands successfully teleported information from one quantum bit to another over a distance of several metres. (The key word here is “information”; data about the quantum state gets teleported, not physical matter.)
More tangible is Canada’s head start in tackling these problems. When universities and institutes in other countries were just beginning to acknowledge quantum computation as a field of study about a dozen years ago, Canada was already going full steam ahead. Waterloo and IQC are leading the way, but there’s also important work being done at the University of Toronto, the University of Calgary, Simon Fraser University and beyond. On the national level, the Canadian Institute for Advanced Research hosts a Quantum Information Science program (with Laflamme acting as director). “Because the rest of the world is just starting to wake up to this, we have an advantage,” says Laflamme. There’s even talk of Canada, and the Waterloo area in particular, as a “Quantum Valley” — a technology hub forming “the epicentre of the next information revolution,” as IQC’s website proclaims.
Scott Aaronson, a computer scientist at the Massachusetts Institute of Technology and a popular science blogger, agrees “Canada is punching way, way above its weight,” noting there’s more quantum research going on in southern Ontario than at MIT. “I think Mike Lazaridis had an incredibly far-sighted vision,” Aaronson says. “Waterloo basically went from being not on the map in these areas [of research] to being the biggest place in the world for quantum computing and information.”
When Martin Laforest was an undergraduate at McGill University, Laflamme dropped by to deliver a lecture and to look for possible IQC recruits. Laforest remembers being immediately hooked on the idea of quantum computation — he “drank the quantum Kool-Aid,” as he puts it. The affable 34-year-old physicist is now the senior manager of scientific outreach at IQC, and one afternoon a few months back he gave me a tour of its many labs. He showed me enormous rooms with shiny canisters of liquid nitrogen; rooms with yellow signs that warn visitors about strong magnetic fields; and rooms where peculiar hums seem to emanate from the ceiling. A “clean room” for preparing materials — apparently you need several months of training before you’re allowed in — is bathed in pinkish-orange light.
The final lab on our tour is built around a forest of polished steel tubes and drums, running a dozen metres in length and branching off this way and that; many component parts, but clearly forming a unified whole. Along the perimeter are gauges, dials and cables. This giant silver beast (which Laforest affectionately calls “the behemoth”) is the Omicron-Oxford multicluster thin film system, a $5-million instrument designed for ultra-high vacuum deposition — a method for producing quantum materials in the form of crystals and thin films, one layer of atoms at a time. These thin films, which can be found in today’s computer chips and microsensors, are expected to play an important role in tomorrow’s quantum devices.
Still, it’s the quantum computer — with the power to blitz through a gazillion calculations in the blink of an eye — that is seen as the big prize, the Holy Grail of quantum information science. And while most researchers say they’re working on it, a company in Burnaby, BC, called D-Wave Systems, claims it has already built a 512-qubit quantum computer — far more impressive than the dozen qubits Laflamme and his colleagues at IQC have been working with. D-Wave has sold one of these megacomputers to NASA, one to Lockheed Martin and, according to Time magazine, one to an unidentified US intelligence agency. The company’s investors include Goldman Sachs, Draper Fisher Jurvetson (which funded Skype and Tesla Motors) and Amazon founder Jeff Bezos. (For more on tech funding in Canada, see “Grow me the money“).
If D-Wave’s quantum computer is the real thing — the matter appears to be open for debate — it should be able to perform 2512 simultaneous calculations. That’s a number much, much larger than the number of particles in the known universe. The catch is that D-Wave’s machines do not seem to be general-purpose computers; rather — as the company admits — they do something called “quantum annealing,” which can be used to solve only a particular class of mathematical problems, known as “discrete combinatorial optimization” problems. (An example is the famous “travelling salesman problem,” in which a sales agent has to visit N cities, spread out at random, in the minimum amount of time.)
“There’s been a lot of controversy about what exactly their devices are doing,” says Steinberg, who serves on the company’s scientific advisory board. While it appears “something quantum is going on,” it’s not clear that D-Wave’s machines “can do a better job than classical algorithms.” Aaronson is more circumspect. “D-Wave has gotten a lot of press by telling the world what it wants to hear,” he says. “Is it actually doing anything better than you could do it with a classical computer? The evidence now is that it isn’t.” Still, D-Wave has built something and investors seem to be willing to gamble on it. Laforest says D-Wave’s contraption is, at the very least, “a great piece of engineering.”
The classical computer evolved over more than a century, from wild-eyed 19th-century visions, through the crucial theoretical work of Alan Turing and John von Neumann in the 1930s and 1940s, to the first general-purpose electronic computer — a room-sized machine known as ENIAC, unveiled in 1946. They’ve been getting faster (and smaller) ever since. Quantum computing, meanwhile, is in its infancy.
“I would say we don’t yet have the quantum-computing analog of the transistor,” says Aaronson. “Right now we’re still at a pre-transistor stage; maybe even a pre-vacuum-tube stage. We’re definitely not at the ENIAC stage.” Steinberg is slightly more optimistic. Quantum computers will eventually be able to solve “problems that we care about,” he says, and will do so faster than any classical computer — but when we’ll get to that stage is up in the air; it could be a decade or more away.
Then there is the question of who will want one. Back in 1943, the chairman of IBM, Thomas Watson, predicted a global demand “for maybe five computers.” His mistake was that he underestimated how versatile computers could be and how many different kinds of problems they could be applied to. “People used to think these were adding machines,” muses Steinberg, motioning toward both the laptop and desktop computers in front of him. “The people who first had the vision that we might use them for, say, word processing, were seen as a little crazy. The idea that we’d use them for images, for sending pictures — that was just beyond the realm of possibility.” Eventually, we began to see how much of the world can be treated as information; that much of what we do boils down to the flipping of 0s and 1s. Something similar may come to pass with quantum computers, Steinberg says. “We know they can manipulate information in certain new ways, faster than our existing computers can. We have to just explore and figure out what the applications of those kinds of manipulations are.”
Back at IQC, Laforest is even more enthusiastic. Beyond the applications being discussed today, he says, are many more waiting to be discovered. “We’re talking about a transformative, disruptive type of technology.” After a brief pause, he adds: “I know I sound like a preacher — but that’s where the potential is. Just like silicon changed everything, quantum information has the potential to change everything.”
Read the original article here.
June 1, 2015
The David Johnston Research + Technology Park is a community where challenges are tackled, advancements are made and new ways…
University of Waterloo student Saajan Kartalks about his recent co-op experience with SAP Labs Waterloo.