Los Alamos National Laboratory
Gary Kliewer, (505) 665-2085, [email protected]
James Rickman, (505) 665-9203, [email protected]
EMBARGOED until Monday, Feb. 17, 9:30 a.m. PST
LOS ALAMOS THEORISTS CORRECT FOR QUANTUM WEIRDNESS
SEATTLE, Feb. 17, 1997 -- Scientists are trying to build a computer that uses the bizarre properties of quantum physics to perform calculations many thousands of times faster than today's supercomputers. Now Los Alamos National Laboratory theorists have brought functional quantum computers closer to reality by demonstrating mathematical methods that correct for the ruinous errors that would creep into quantum calculations.
Los Alamos theorist Raymond Laflamme will discuss "Threshold Accuracy for Quantum Computation" today at the quantum computation session of the American Association for the Advancement of Science conference.
A quantum computer could unravel within seconds the encryption codes that today are unbreakable due to the phenomenal amount of computing power required. However, no one has built a working quantum computer yet, and the first one may be decades away. It is dauntingly difficult to build a machine that manipulates individual atoms to perform many calculations at once by taking advantage of quantum physics. Some scientists argue that quantum computers could never work because quantum states are so fragile there is no way to correct for errors. Error correction strategies used in today's computers are useless for quantum calculations.
But if a quantum computer can be built, Laflamme and his Los Alamos colleagues Emanuel Knill and Wojciech Zurek have found error correction strategies that will make the calculations trustworthy.
"As recently as two years ago, the main problem for quantum computers was an inability to correct errors," Laflamme said. "Last year, quantum error-correcting codes were discovered, and they protect quantum calculations so that a limited number of errors won't kill the operation, even when errors occur in the encoding process itself."
"Still there were limits," Laflamme said, "because residual errors multiply and eventually destroy confidence in the calculation."
The solution, in simplest terms, is to use what computers do best -- repetitive processing -- to reduce the probability of errors. In Los Alamos' scheme, every encoded quantum bit, or qubit, is checked for errors, corrected, then multiplied five times. Those five qubits also get checked for errors then corrected and multiplied, and so on. Knowing how many steps a particular calculation takes, the theorists know the number of checks needed to ensure the calculation's accuracy.
"Now we know a quantum computer does not have to be perfect. If we can achieve a threshold level of accuracy, there is no limit to the amount of calculations possible. There is no fundamental limit to quantum computing," Laflamme said.
Embassies that transmit sensitive messages, scientists managing the nuclear weapons stockpile, designers of browsers for the World Wide Web and your local bank all are watching to see if quantum computing is a dream of theorists only, or if it can become practical. A quantum computer would excel at factoring numbers, a process inherently unmanageable by conventional computers and therefore used for encryption of confidential information.
Interest in quantum computing has grown as both practical applications and potential paths to a working machine have appeared in the last two years, but everyone in the field admits a quantum number cruncher is still years away. With more researchers involved, the odds improve that someone will come up with a winning design. "But the 'Bill Gates' of quantum computing has probably not been born yet," said Laflamme.
Los Alamos National Laboratory is operated by the University of California for the U.S. Department of Energy. -30-
RSS