Quantum computing feasible if errors can be corrected
Copyright 1998 Nando.net
Copyright 1998 Scripps-McClatchy Western
ALBUQUERQUE, N.M. (January 17, 1998 00:19 a.m. EST
http://www.nando.net)
- Scientists at Los Alamos National Laboratory say the still-visionary
but powerful concept of "quantum computing" is feasible if designers
can
build in ways to automatically correct for anticipated errors.
Such advanced computers, which would use atoms to hold and store
information, will need to store that information redundantly, or many
times, the scientists say.
And they will have to use a system to constantly check and correct
errors to which the high-speed atomic computer itself would be prone.
The research, published in Friday's issue of the journal Science,
"really opens the door and makes a broad road for the field of quantum
computing to follow," said Los Alamos physicist Raymond Laflamme, one
of
three authors who did the research.
Laflamme, a physicist, collaborated with Los Alamos colleagues Emanuel
Knill, a mathematician, and Wojciech H. Zurek, a physicist.
Their Science paper, called "Resilient Quantum Computation," states
that
"accurate quantum computation is possible provided that the error per
operation is below a threshold value."
Unlike today's computers, which use electronic gates to represent a one
or a zero on silicon wafers, quantum computing would use atoms to store
computer data representing the information atomically as ones, zeros,
and even combinations of ones and zeros.
Laflamme said that would greatly magnify the amount of information and
the speed with which information can be processed. However, "(atomic)
states are very fragile" and the backup error-checking and restoration
system is essential to maintaining accurate information, he said.
Quantum computing, still in its conceptual infancy, would be an
enormous
leap in the power of machines to handle massive numbers and solve very
complex problems easily.
Laflamme said such machines would be many times more powerful than the
fastest supercomputer, currently at Sandia National Laboratories in
Albuquerque.
Scientists last fall used the Sandia machine, built by the Intel Corp.,
to set a world record of 1.34 trillion calculations per second, said
Art
Hale, Sandia's computer-science director.
Hale, who had yet to see the Los Alamos paper, said the work is "headed
in the right direction" because existing computer technology has
built-in limits on size and speed.
"At some point we can't make it (current computers) any smaller or
faster," he said. "The atomization or quantum effort will be the next
big step. Obviously, we are going to have to tackle this problem
head-on, and it sounds like they have made a significant contribution."
Also at Sandia, Jim Tomkins said that while quantum computers remain "a
ways off," they are tantalizing to many scientists.
Tomkins, who oversees Sandia's supercomputer, said quantum designs
would
offer "high, dense storage and speed capabilities."
"You can store a huge amount of information and very densely when you
think of the huge number of atoms available," he said.
Laflamme said that characteristic is also what makes the error-checking
solution feasible. He said because of the vast capacity of a quantum
system, the information redundancy and error-checking "consume only a
fraction of the number of bits (information storage units) available."
While scientists dream of quantum machines that could virtually
instantly juggle monster numbers and perhaps accurately model big
science problems like global warming, Laflamme said there are practical
social applications.
One, he said, would be guaranteeing security in electronic financial
transactions by using a code based on a key with such long numbers that
it would be virtually impossible to crack.
Encryption, such as coding of classified government information, is one
of the big reasons the Los Alamos scientists are working on the
problem.
Los Alamos is the nation's premier nuclear-weapons laboratory. Sandia
has prime responsibility for engineering nuclear weapons.
Tomkins said Sandia, which helped spawn the parallel computing
revolution that is the basis of its supercomputer, said Sandia
continues
to focus on "more near-term computer advances."
Many computer scientists have doubted that quantum computing is even
possible because it will be so prone to routine errors caused by the
inherent inability to maintain information integrity, atom by atom.
Atoms easily lose and gain electrical charge - and that introduces a
major source for error.
That's why the Los Alamos research is considered important in laying a
foundation in the field.
By LAWRENCE SPOHN, The Albuquerque Tribune
This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 13:01:36 PDT