IBM's Watson supercomputer uses parallel computing to connect highly specialized pieces of software to work simultaneously to answer questions posed by the Jeopardy! game show. Watson uses its huge database to identify patterns of words and then searches for answers that fit the question.
In the future, a Watson-like system could perform medical diagnoses, writes Yale University computer science professor David Gelernter. For example, if a doctor sees an unusual or unfamiliar case, they could submit it to the supercomputer's database. The system would find similar cases and offer information or advice.
A theoretical WikiWatson would contain information submitted by experts from all over the world. Gelernter also notes that the system's performance would constantly improve as more information is added.
IBM has been conducting parallel artificial intelligence (AI) research since the early 1980s, starting with a program that could prove theorems in high school geometry. Although Watson has not passed the Turing test, in which a computer system tricks a human into believing the system also is human, it is a major step forward for AI technology, Gelernter says.
From The Wall Street Journal
View Full Article
Abstracts Copyright © 2011 Information Inc. , Bethesda, Maryland, USA
No entries found