IBM releases source code

TalkBack 9 of 11:

Next »

« Previous

• Thread View
• Flat View

Granted there is an accuracy issue...

but the English language has roughly 200,000 words (Oxford English Dictionary, Oxford American Dictionary), most people have a usable vocabulary of roughly 20,000 words, some have more. Since English is not a tonal language (as are dialects of Chinese, for example), there is a built in accuracy problem. WHen last checked, the best (or most accurate) was Dragon Naturally Speaking, with a 95% accuracy but still limited to roughly 20,000 words (how much drive space do you want to give an application && database?).

If you have an accent (Southern, for example), this may confuset the language parcer.

On a side note, a device used by the U.S. military to do translation (audible) for farsi (sp?) in Iraq proved to be fairly accurae, though within a fairly limited context. If you are expecting some natural language parcer to handle some 200,000 words, verbal speeds, context parsing, (...) you mayt be expecting a bit muh (for now).

Other languages are easier to get correct for speech processing, as there is not an accent issue!

Posted by: B.O.F.H.   Posted on: 09/27/04 You are currently: a Guest | Members login | Terms of Use

Alert moderator to an offensive message

Subscribe to this discussion via Email or RSS

SponsoredWhite Papers, Webcasts, and Downloads

• Get top-ranked Novell support for Red Hat at 50% less Novell A simplified IT environment isn't just less complex, it's more reliable. ... Download Now
• VMware Infrastructure: A Guide to Bottom-Line Benefits VMware Frustrated by the costs of maintain ever larger data centers?or building ... Download Now
• The True Costs of Virtual Server Solutions VMware In an economic environment that is repeatedly heralding the message "do ... Download Now