Quantum computing has long been a wacky, borderline fictional, mostly theoretical domain of physics reserved for highly speculative conversation. This is because quantum mechanics, or particle physics as it’s also called, makes some claims completely void of common sense. Particle physicists believe that a subatomic particle called a neutrino can pass through the entire Earth without slowing down, and that particles can be in two different states at the same time, and even that two particles can be entangled in such a way that their properties will match across any distance (imagine if flipping a light switch in Kansas caused a light switch on Saturn to flip as well). Various governments have poured money into the exploration of these theories—a giant sub-atomic roller rink was built in Geneva, Switzerland to test many of them resulting in the discovery of the Higgs Boson. But there hasn’t been much use for these theories in practical application. That is until the concept of quantum computing came about.
If a particle can be in two states at once, then perhaps this could be used to speed up computation by incredible amounts. Traditional bits in computers can either be on (1) or off (0), but quantum bit, or qubits, can be both on and off at the same time (called superposition), allowing them to perform parallel computations at once. Make a device that runs with qubits as the base system and you’ve got a quantum computer. The ability to be both on and off simultaneously allows quantum computers to use a process called annealing and makes a computer extraordinarily faster as it’s able to process all scenarios at once. A quantum computer with just 300 qubits could run more calculations in an instant than there are particles in the whole universe. But all of this is just in theory.
As the theories about quantum computing grew and they became entangled with increased media attention and speculation about their capabilities, the idea of quantum computing morphed into miracle computing. It was believed if quantum computers existed, they could cure disease, power artificially intelligent robots, make time machines function, drive cars, and solve the problems of global warming.
While all the speculation was brewing and more scientists and researchers wrote papers on the matter, one company started to build quantum computers. In 2011, a Canadian company D-Wave (backed by the CIA and individuals like Jeff Bezos of Amazon) sold its first machine to defense contractor Lockheed Martin. In early May, D-Wave sold its second machine, the 439-qubit D-Wave Two, to the Quantum Artificial Intelligence Lab for $15 million. The lab, which is backed by NASA, Google, and the Universities Space Research Association (USRA) will use the device to make advances in machine learning, a field of computer science where computers become more adept at solving problems with the more experience they have.
Research Catherine McGeoch, a professor of computer science at Amherst College, but theory into practice and tested the D-Wave prototype to see if it really was a quantum leap forward. Her findings conclude that the device is fast, but only at specific tasks. “On the largest problem sizes tested, the V5 chip found optimal solutions in less than half a second, while the best software solver, CPLEX, needed 30 minutes to find all optimal solutions,” McGeoch writes in the conclusions section of her academic paper, where CPLEX is a conventional software solver and V5 is the chip in the D-Wave prototype. They received a second V6 chip after most of the study had finished, but they decided to test it anyway, concluding “V6 is three to five times faster than V5″ and “preliminary results suggest that… the hardware can find optimal solutions around 10,000 times faster than CPLEX.”
But these incredible numbers can be a little misleading. This doesn’t say that the D-Wave Two is generally 3,600 to 10,000 times faster than a conventional computer, rather that it solved a specific problem that much faster than the current standard solver CPLEX. As McGeoch told the New Yorker after the many media organizations stated the quantum computer was 3,600 times faster, “the 3,600 number does not give any information about comparative performance of the two types of platforms. It was never intended to.”
Another misleading detail is that the baseline machines that the D-Wave Two was being compared against are simple desktop machines that cost only $1,200. The D-Wave machine wasn’t being compared to state-of-the-art supercomputers, but with something you could more or less pick up at Best Buy. For the cost of one D-Wave Two, you could buy 12,500 of the traditional machines. This doesn’t exactly seem like a fair comparison, and it doesn’t even account for the fact that there need to be incredible conditions to make the D-Wave Two run. Because the machine’s chip requires a near absence of electrical resistivity to function called superconductivity, the machine must be supercooled to nearly absolute zero.
Furthermore, there’s even some doubt as to whether D-Wave’s machines are actually quantum computers. It’s very difficult to tell if a device is actually using a process called quantum tunneling or if a similar effect is being achieved through normal thermal fluctuations. McGeoch even admitted she wasn’t sure how the machine actually operated and simply deferred to previous research that said the D-Wave machine is “at least a little quantum mechanical.”
All things considered, it seems we only have an incredibly expensive machine that looks like a quantum computer. Even at its best, a true quantum computer isn’t the magical solution we might be looking for. Due to the way quantum devices are structured, they may excel at solving specific problems that require multiple calculations simultaneously (like determining how to seat guests at the dinner table so people who dislike each other aren’t placed together), they fall short at doing other computational tasks like running Photoshop or Microsoft Word or browsing Facebook. While this is definitely a simplification, it assists the point that with the current state of the technology, quantum computers will have to be coupled with traditional computers to comprehensively perform tasks. In Google’s announcement surrounding purchasing the D-Wave Two, Director of Engineering Hartmut Neven admitted that in trying to better understand machine learning “we’ve learned some useful principles: e.g., you get the best results not with pure quantum computing, but by mixing quantum and classical computing.”
Even though quantum mechanical devices may be here to the tune of $15 million, there still will be years of research, development, and debate to determine if the technology is the miracle device science fiction hopes it is.
“Launching the Quantum Artificial Intelligence Lab”, Google
Adrian Cho, “Controversial Computer Is at Least a Little Quantum Mechanical”, Science
Gary Marcus, “A Quantum Leap In Computing?”, The New Yorker
Charles Choi, “Google and NASA Launch Quantum Computing AI Lab”, MIT Technology Review
Catherine C. McGeoch, “Experimental Evaluation of an Adiabiatic Quantum System for Combinatorial Optimization”
John Naughton, “Is computing speed set to make a quantum leap?” The Guardian