Publication

Using out-of-language data to improve an under-resourced speech recognizer