Computer spellcheckers work fine unless you accidentally type in a correctly
spelled word that is incorrect in meaning, such as keying in 鈥渟ight鈥 instead of
鈥渟ite鈥. But now Klaus Truemper of the University of Texas in Dallas has
developed a spellchecker that can weed out these errors by learning how the user
composes text. Like early voice-recognition systems, it uses statistical
analysis to try and decide what the text should mean and whether the correct
words have been used. Early tests show the system suggests the correct word for
an error 96 per cent of the time.
More from New Scientist
Explore the latest news, articles and features
Technology
Test moderators use AI-generated writing to judge literacy standards
News
Mind
THC in cannabis may reduce nightmares in people with PTSD
News
Environment
Fishers detonate a bomb every hour on one Indonesian coral reef
News
Environment
How bad is the UK drought and how long will it continue?
News
Popular articles
Trending New Scientist articles