Sakriani Sakti

Incorporating Knowledge Sources into Statistical Speech Recognition