Skip to content

The Very Strange—And Fascinating— Ideas Behind Quantum Computing

2016 July 31
by Greg Satell

In 1952, Remington Rand’s UNIVAC computer debuted on CBS to forecast the 1952 election as early results came in. By 8:30, the “electronic brain” was predicting a landslide, with Eisenhower taking 438 electoral votes to Stevenson’s 93. The CBS brass scoffed at the unlikely result, but by the end of the night UNIVAC proved to be uncannily accurate.

It was that night that the era of digital computing truly began and it was a big blow to IBM, the leader in punch card calculators at the time. It’s Research division, however, was already working on more advanced digital technology. In 1964, it launched its System 360 and dominated the industry for the next two decades.

Today, we’ve reached a similar inflection point. Moore’s law, the paradigm which has driven computing for half a century will reach its limits in about five years. And much like back in the 1950’s, IBM has been working on a new quantum computer that may dominate the industry for decades to come. If that sounds unlikely, wait till you hear the ideas behind it.

A 90 Year-Old Argument

In the early 20th century, one of the fundamental assumptions was an idea, sometimes known as Laplace’s demon, that the universe was perfectly deterministic. In other words, if you knew the precise location and momentum of every particle in the universe, you could calculate all of their past and future values. Every effect has a cause, or so it was thought.

Yet by the 1920’s, many began to question that idea and the issue came to a head in a series of debates between Albert Einstein and Niels Bohr. It was then that Einstein famously said, “God does not play dice with the universe.” To which Bohr cleverly retorted, “Einstein, stop telling God what to do!”

At issue were two ideas in particular. The first was quantum superposition, or the principle that particles can take on an almost ghostly combination of many states at the same time. The second is quantum entanglement, which says that it is possible for one particle with unpredictable behavior to allow you to perfectly predict the behavior of another one.

These are hard ideas to accept because they run counter to what we experience in normal life. Everyday physical objects don’t simply appear and disappear, or start jetting off in one direction for no particular reason. Einstein, who certainly did not lack imagination, could never accept them and devised an experiment, called the EPR paradox to disprove them.

Yet it is exactly these ideas that IBM is betting on now. To help me wrap my head around it all, I spent several hours talking to Charlie Bennett, an IBM Fellow considered to be one of the founders of quantum information theory.

A Geek Before Geeks Were Cool

Growing up in the quiet Westchester village of Croton-on-Hudson, about a half hour from IBM’s headquarters in Armonk NY, Bennett was, as he put it to me, “a geek before geeks were cool.” While other teenage boys were riding bikes and playing baseball, he usually had his head buried in a copy of Scientific American, wrapping himself in its world of crazy ideas.

And in the 1950’s, there were more than enough fantastical discoveries to go around. Many things we take for granted today, like computers that work as “electronic brains” and nuclear energy, were novel back then and just beginning to be understood. However, what enthralled him the most at the time was Watson and Crick’s discovery of the structure of DNA.

So when he went of to college at Brandeis, Bennett was determined to become a biochemist. Unfortunately, the university didn’t offer that as a major, so he got his degree in chemistry and then went to Harvard to study molecular dynamics under David Turnbull and Berni Alder, two giants in the field.

Yet even that heady work was unable to quench his curiosity, so Bennett branched out. He took a course about mathematical logic and the theory of computing, which introduced him to the ideas of Kurt Gödel and Alan Turing, while at the same time working as a teaching assistant for James Watson, who won the Nobel prize for the discovery of the structure and function of DNA just a few years earlier.

Oddly, he found his two extracurricular activities to be two sides of the same coin, with the DNA transcription machinery eerily similar to a Turing’s ideas about a universal computer. It was that insight—that the world of computation could be more than a sequence of ones and zeros—that set him on his course. He began to see strange forms of computation almost everywhere he looked.

A Witches Brew Of Crazy Ideas

As a graduate student, Bennett went to see a talk by an IBM scientist named Rolf Landauer and learned about his principle that if bits are not erased, then energy can be conserved. With his background in chemistry, Bennett was able to further Landauer’s work and make important breakthroughs in reversible computing. Bennett was soon thoroughly hooked on computing—and on IBM.

Although he had planned on a career in academia, he found that, “being at the Yorktown lab gave me the opportunity, within one building, to interact with physicists, engineers, and computer scientists and learn about their fields. Over the subsequent 44 years, I’ve had the freedom to think about what I wanted, and to visit and collaborate with scientists at universities and laboratories all over the world.”

It was that ability to explore new horizons without limits that drove Bennett’s work. For example, his friend Stephen Wiesner came up with the idea of quantum money that, because of the rules of quantum mechanics, would be impossible to counterfeit. It was the first time someone had a concrete plan to use quantum mechanics for informational purposes.

Weisner’s insight led Bennett, along with Gilles Brassard, to develop the concept of quantum cryptography, which has a similar logic to it. Anybody attempting to eavesdrop on a message encrypted quantumly would destroy the message. These were breakthrough ideas, but what came next was even more impressive.

Einstein’s Last Stand

As noted above, Einstein could never bring himself to accept quantum mechanics, especially entanglement, because he thought that such “spooky action at a distance” violated the laws of physics. How could observing a particle in one place tell you about a particle in another place, without affecting it in some way?

Einstein felt so strongly about the idea that he devised an experiment, called the EPR paradox, to finally prove or disprove the concept. In a nutshell, he proposed to test the principle of entanglement by using one particle to predict the behavior of another one. John Bell showed this could be indeed be done and other scientists verified his results in a lab a few years later.

Armed with their insights quantum cryptography, Bennett and Brassard, along with a number of colleagues, took Bell’s work a step further in the famous quantum teleportation experiment carried out in 1993, which not only made clear that was Einstein wrong, but that quantum entanglement could actually be far more useful than anyone had dreamed.

Yet Bennett still had his sights set on an even bigger prize—using quantum states to compute, rather than just relay, information. What he was proposing seem almost incomprehensible at the time—a computer based on quantum states potentially millions of times more powerful than conventional technology. In 1993, he wrote down four laws that would guide the field.

A New Quantum Universe Of Computing

To understand how a quantum computer works, we first have to think about how a classical computer, sometimes known as a Turing machine, works. In essence, today’s computers transform long series of ones and zeros — called bits — into logical statements and functions according to a set of rules called Boolean logic.

Now, ordinarily, this would be an incredibly foolish way to go about things because you need a lot of ones and zeros to explain anything, but today’s computers can do literally billions of calculations per second. So at this point, we are able to communicate with machines in a fairly reasonable way, such as typing on a keyboard or even talking into a microphone.

To get an understanding of how this works, let’s look at a character. Eight bits gives us 28, or 256, possible combinations, which is plenty of space to accommodate letters, numbers, punctuation and other symbols. With processors able to handle billions of bits per second, we can get quite a lot done even with basic, everyday machines.

The math of quantum computers works in a somewhat similar way, except because of superposition and entanglement, instead of combinations, it produces “states.” These states do not conform to any physical reality we would be familiar with, but roughly represent separate dimensions in which a quantum calculation may take place.

So an eight quantum bit (or qubit) computer can be in a superposition of 256 different states (or dimensions), while a 300 qubit computer can be simultaneously doing more calculations than there are atoms in the universe.

There is, however, a problem. These “states” represent only possibilities. To get a quantum computer to focus on a single concrete answer is a very complicated business. When the quantum computer is being used to answer a quantum question, such as how the human body interacts with a new drug, this focusing happens automatically.  But in other cases, such as when a quantum computer is used to answer a classical question, major difficulties arise.

The potential of quantum computing is immense, so computer scientists at IBM and elsewhere are working feverishly to smooth out the kinks—and making impressive progress. IBM has also made a prototype quantum computer available in the cloud, where even college students can learn how to program it.

We Are Entering A New Quantum Era

The ideas surrounding quantum computing are so strange that I must confess that while talking to Dr. Bennett, I sometimes wondered whether I had somehow wandered into a late night dorm room discussion that had gone on too long. As the legendary physicist Richard Feynman confessed, the ideas behind quantum mechanics are pretty hard to accept.

 

 

Yet as Feynman also pointed out, these are truths that we will have to accept, because they are truths inherent to the universe we live in. They are part of what I call the visceral abstract— unlikely ideas that violate our basic notions of common sense, but nevertheless play an important part in our lives.

We can, for example, deny Einstein’s notions about the relativity of time and space, but if our GPS navigators are not calibrated according to his equations, we’re going to have a hard time getting to where we’re going. We can protest all we want that it doesn’t make any sense, but the universe doesn’t give us a vote.

That’s what’s amazing about people like Charlie Bennett. Where most people would say, “Gee, that’s weird,” he sees a system of rules that he can exploit to create things few others could ever imagine, almost as if he was playing the George Clooney character in Ocean’s 11. But instead of scamming a casino, he’s gaming the universe for our benefit.

“Charlie is one of the deepest thinkers I know,” says IBM’s Heike Riel. “Today we can see that those theoretical concepts have come into fruition. We are on the path to a truly practical quantum computer, which, when it’s built, will be one of the greatest milestones not just for the IBM company, but in the history of information technology.”

So we now find ourselves in something much like those innocent days before 1952, when few could imagine something like UNIVAC could outsmart a team of human experts. In a decade or two, we’ll most likely have to explain to a new generation what it was like to live in a world without quantum computers, before the new era began.

– Greg

3 Responses leave one →
  1. July 31, 2016

    Fascinating.

    The best I can do is to relate back to my work on the feasibility of creating a thinking machine back in the 1960s.

    At the time the digital computing people thought they were leading the field.

    I thought that the time taken to process the information in that way was too slow and a highly complex task which would take a long time to make practical.

    Instead I proposed a learning network based upon the kind of network to be seen between cells in the brain and I wanted it all to be based upon pattern recognition which I defined as knowledge gained.

    Today this kind of research is making some progress I believe though I am no long in touch.

    But I also did information theory. It is clear from that that an item which can have ten states contains far more information than an item which can have two states.

    A neuron which can make a choice of ten different connections is similarly the possessor of much more information than one which has a choice of two.

    A network of such neurons can be in trillions of such states even if it involves just a hundred or so neurons. Maybe 100 to the power ten. That is a lot of information.

    If that quantum computer can have 256 states, not ten, then as you say, the number of states and the amount of information which can be handled may well be greater than the total number of atoms in the universe.

    What is 100 to the power 256?

    And if the array is able to process information in the kind of way that a human mind can…but there I have to admit I am unable to understand how that proposed processing model would work or might be programmed.

    Looking forward to the future if I live that long

    Edward

  2. July 31, 2016

    Thanks for sharing your thoughts Edward.

    – Greg

  3. Mary permalink
    February 12, 2017

    I enjoyed very much the article. And suddenly thoght and what about the networks? Hoe are they going to endure the kind of content and communucations that can be produced with such computing capacity?
    Thank you!

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS