It's been about a month since I didn't tell you anything about quantum computing. Our topic for today is not about hardware, but more specifically about software for pattern recognition and image processing.
A physicist at the University of British Columbia has come up with an algorithm that proves that quantum computers would be faster at finding patterns. "Finding and recognizing [a linear] pattern can be accomplished much faster on a quantum computer than on a classical one," said Ralf Schützhold, a researcher at the University of British Columbia.
The algorithm would allow quantum computers to detect an 8-by-8 grid of alternating black and white squares set in an array of 640 otherwise randomly distributed squares.
This seemingly simple task takes a classical computer about 6,000 steps because it would have to compare each square to every other square, one at a time.
A quantum computer, however, can examine all of the possible solutions to a problem at the same time, in this case comparing all the squares to each other at once.
Sounds promising, isn't? This researcher is pretty confident anyway. Listen to him.
Pattern finding is a key component of speech, face, and handwriting recognition programs, and of software that sorts seismographs and other large sets of scientific data, said Schützhold. The exponential speed-up promised by quantum computers might enable us to attack problems that would take classical computers "longer than the age of the universe" to solve, he said.
To see some practical applications, you'll have to wait for about twenty years. But if you want to read Schützhold's paper today, here it is: "Pattern recognition on a quantum computer."
If you're not a mathematician, forget it.
Source: Eric Smalley, Technology Research News, September 4-11, 2002
5:39:55 PM
Permalink
|
|