Improved hindi broadcast ASR by adapting the language model and pronunciation model using a priori syntactic and morphophonemic knowledge.

Preethi Jyothi, Mark Hasegawa-Johnson

Browse the full Interspeech paper archive.