On cross-language experiments and data-driven units for ALISP (Automatic Language Independent Speech Processing)

The authors discuss the cross-language use of phonemic models, in the context of their research goal: Automatic Language Independent Speech Processing (ALISP). A method is proposed and demonstrated for creating adequate representations of target vocabulary words through sub word models, given minimal training data. Also, the topic of data-driven determination of language-independent acoustic segments for speech recognition is addressed.