Features

Lev Levitin and Tommaso Toffoli of Boston University suggested something similar in September 2007. Quantum computing is very sensitive to thermal noise, so any quantum computer designed to work reasonably fast will require extra qubits for redundancy and error correction. These requirements may cancel out the benefits of quantum computing, they argue.

Scott Aaronson, assistant professor of electrical engineering and computer science at MIT, has taken on board the difficulties, and goes back to the double-slit experiment for an explanation of what a real quantum computer might be: "The idea of quantum computing is to set up a massive double-slit experiment with exponentially many paths, and to try to arrange things so that the paths leading to wrong answers interfere destructively and cancel each other out, while the paths leading to right answers interfere constructively and are therefore observed with high probability."

A commercial system

Quantum computing doesn't lend itself to commercial funding. Even with the huge possible benefits, the pay-off is so far in the future that it isn't a good way for companies to spend too much of their money.

Companies such as IBM and HP have only dabbled in quantum computing. "A company that intends to be involved in the computational business a couple of decades from now should definitely be looking at quantum computation, but it is betraying its stockholders if it is currently spending more than a million dollars a year working on this area," says Stan Williams, director of basic research

ADVERTISEMENT

for HP and founder of the company's quantum science research group.

Despite this, there is a commercial quantum computing company. D-Wave Systems, based in Canada, has raised $44 million in venture capital to fund its plan to deliver supercooled quantum computers to discerning (and rich) computer rooms in the next couple of years.

The company is the brainchild of maverick scientist Geordie Rose. Its approach uses adiabatic quantum computing (AQC), an alternative quantum computing technique described in a paper co-authored by Seth Lloyd.

Rose states on his website that the big benefit of AQCs is they don't need to be shut off from the world. "AQCs can be run in the presence of noise without quantum error correction, and still provide optimal scaling. In fact the thermal noise actually helps."

An adiabatic quantum computer is designed to find the solution for one particular problem. When an AQC interacts with the outside world, its quantum systems settle to the state where they have the lowest energy, like a ball rolling downhill. An AQC is designed so that - for a given set of inputs - its lowest energy state is the answer to the problem.

A system's overall quantum state is called the Hamiltonian. This complex mathematical function describes all the possible quantum values the system could have. It predicts the way the system will evolve over time. It dates back to 1833, when Irish mathematician William Rowan Hamilton rewrote classical mechanics, (including Newton's familiar laws of motion), as a set of equations. In classical physics, the Hamiltonian was a mathematical curiosity, showing underlying harmony and symmetry, but not very useful. In quantum mechanics, it turned out to be the only sensible way to express what is going on.

It is possible to define a system with a Hamiltonian such that the minimum energy state is the answer to the problem. An adiabatic quantum computer starts in a well-understood Hamiltonian, with the particles in their minimum energy state. Then the system is gradually changed to the new Hamiltonian. If the particles stay in their minimum energy states, the new Hamiltonian is also solved.