Sixty years ago, the idea of mother earth’s inhabitants being so ruled by computers that they would have to band together to solve a problem like Y2K was barely a vision of the nascent members of the computer field or even science fiction writers.
Though computing power has advanced at incredible speeds over the last quarter century, it has limitations. When you get right down to it, the problem with computers is that they can only do one thing at a time. There is no question they do it quickly, often more quickly than we can ask them to perform a task, but nevertheless, at any given moment there is only one computation going on.
Today there is a growing field of computing so revolutionary that many of the traditional laws of physics which we treat with reverence are getting tossed on their sides.
The field is quantum computing, where computers exploit quantum mechanical interactions to potentially create computers so fast much of today’s cryptography would eminently crackable.
In a Scientific American article a few years back two members of the quantum computing community, Neil Gershenfeld and Isaac Chuang, wrote about how a quantum computer could theoretically factor a 400 digit number, an accomplishment needed to crack some security codes, in a year. On the other hand, today’s fastest supercomputers would take billions of years to perform the same feat, they wrote. Stunning computational power indeed.
Members of this growing, yet relatively young, community see quantum computing today as being evolutionarily at the same point traditional computing was half a century or so ago.
The theories have been proven and the lab work is just starting to take shape. We are decades away from a truly functional product and are even unsure of what we would ultimately do with a quantum computer once one is built, though cryptography will probably the first area exploited by this technology.
“Most of the applications I expect haven’t even been conceived yet, [since] it is fundamentally a different way of computing,” said Mike Mosca, a professor of mathematics at St. Jerome’s University at the University of Waterloo in Waterloo, Ont.
understanding it all
Today’s two bit registers have four possible states. Two ones, two zeros, a zero and a one, and a one and a zero.
“The quantum computer register can be all four at the same time, so enormous parallelism, which leads to exponential improvement in the power of computing, [is possible],” said Nabil Amer, manager of the physics of information department at IBM in Yorktown, NY. This ability to be in multiple states at once is called superposition.
“[This] sounds terrible, since usually you would think the last thing I want in my computer is to not know where a bit is or whether it is a one or a zero but you can use that in a more controlled way where a given bit might be, at some level, a zero and also a one at the same time,” said Aephraim Steinberg, professor of physics at the University of Toronto.
The accompanying information box helps explain how a single photon of light can be in two places at once (essentially being a zero and one simultaneously) and also how a quantum computer would have an exponential increase in power.
But making that computer is no easy task. The ideal quantum computer is completely isolated from its surroundings. This, in itself, creates huge problems. How do you ask a quantum computer to perform a task if asking it a question destroys its quantum state?
The problem is that almost any interaction with a quantum system constitutes a measurement. This is akin to measuring the temperature of a swimming pool. The simple act of putting a thermometer into a pool changes the temperature. The thermometer takes heat from the pool’s water to get an accurate measurement and in doing so actually changes, albeit infinitesimally, the pool’s temperature. In a quantum situation the act of measuring the system collapses the superposition of the quantum mechanical state into one that is a single definite state. This is called decoherence. Though you can measure it, the measurement may not be accurate, as is the case of the thermometer’s measurement of the pool.
Thus a quantum computer must have its inner workings isolated from its surroundings but still be accessible to load, execute and read calculations. Quantum ‘error correction codes’ have been found. This allows for the possibility of building quantum computers robust enough to deal with these inherent imperfections.
“We need to find a system that is so well isolated that if we put it into some uncertain state it keeps that information secret, it doesn’t start broadcasting it and accidentally get measured,” Steinberg said.
“It is a very funny thing in that you kind of have to ask the question but cover your own ears while you are asking it,” he joked.
We are years away from solving this conundrum, according to those spoken to.
“The reason we don’t have a good system is that you need something that doesn’t talk too much to the outside world but still lets you ask a question and also has strong enough interactions so that you can make something like a transistor inside it,” Steinberg explained.
Much experimentation is going on, mostly in North America and Europe, in how to build this magical machine. Theories abound but there is considerable work being done using nuclear magnetic resonance techniques to create the necessary isolation for quantum computation. Others are looking into the world of optics to solve the fundamental problems of isolation.
IBM has succeeded in building seven qubit computers at its labs, according to Amer. In an earlier qubit computer, radio-frequency pulses applied to chloroform molecules cause the hydrogen nucleus to rotate between the zero state and the superposition of zero and one, according to the company.
But scalability is the big issue since quantum computers would need around 1,000 qubits to start solving meaningful problems.
“Since we don’t have any show stoppers on the horizon the issue becomes scalability,” Amer said.
“What we are pushing for…is a solid state implementation. “
“Physics tells me I can do it, it boils down to a technological challenge,” he added.
“We are at the juncture where all these paths are out there and I think the healthiest approach is to sort of let 1,000 flowers bloom and then let the Darwinian approach take care of it,” he concluded.
Mosca agrees. “I would expect we are going to proceed bit by bit, that is the only thing you can sort of plan, but inevitably we will be surprised somewhere and advance faster on some fronts,”
“We are looking for a breakthrough (in technology) right now, ” Steinberg added. “If there is an idea out there, I think we will find it in the next 20 years,” he predicted.