![]()
Visiting guest. Why not sign in?
|
|
Dead-End for Speech Recognition Diverging Opinions | |
Dead-End for Speech Recognition
In contrast to two previous articles posted earlier this week (first, second), a new ZDNN feature entitled Voice Recognition: Another Dead End claims that voice technology is not getting too far at all. The key point of the article is that we are so close to getting speech software reliable, which is driving money and effort into the area. However, this seems to be making no difference... First, the naivety of spell checkers is underlined, since they work on a purely reactive basis. Secondly, translators are also criticized, based on the performance of the Google translator. Finally, the article asks how voice recognition can succeede at all, when the underlying understanding is so bad! So, in brief, Natural Language Processing before Voice Recognition; get your priorities right. |
|
Voice Over IP
Vaguely related: |
|