New research shows that automatic speech recognition and speech translation systems also improve when training data includes multiple instances of a given name.

The post In Tech-Assisted Interpreting, the Tricky Problem of Non-English ‘Person Names’ appeared first on Slator.



Source link

X
X
X